Research Paper

  Topic: Research paper on the movie: The Eternal Sunshine of the Spotless Mind.

Primary source: Watch movie

Sources: Minimum five (5) separate secondary sources:

1. No more than one (1) internet source (Databases are okay)

2. At least four (4) sources procured through our collegiate library resources: databases, the library catalog, and Mobius.

3. Secondary sources should be EXPLICITLY about the primary source.

 Generally, sources should be academic and scholarly in nature. Examples of scholarly sources include:

1. –scholarly/academic journals articles

2. –books

3.  –major daily newspapers (Washington Post, L.A. Times, etc.)

MICHAEL J. MEYER

Reflections on Comic Reconciliations: Ethics, Memory,
and Anxious Happy Endings

Comedies of remarriage typically display a char-
acteristic view of the reconciliation of troubled
friendships. My preliminary thesis is that the “re-
marriage reconciliations” at the heart of comedies
of remarriage ground an important and yet under-
appreciated aesthetic category that I call “anxious
happy endings.” Remarriage comedies also high-
light the often overlooked question: are there eth-
ical reasons to try to hold on to certain memories?
I investigate a crucial aspect of this question by
considering the recent film Eternal Sunshine of the
Spotless Mind (Michel Gondry, 2004), which pro-
vides an informed study of ethics and memory.

i. an ethics of thick relations and memory

An insightful analysis of ethics and memory has
been provided by Avishai Margalit.1 I begin this in-
vestigation with his version of a now standard divi-
sion: “the distinction between ethics and morality.
. . . [T]his in turn is based on a distinction between
two types of human relations: thick and thin ones.
Thick relations are grounded in attributes such as
parent, friend, lover. . . . [T]hin relations, on the
other hand, are backed by the attribute of being
human.”2 My present interest is not with morality
but instead with an ethics of thick relations and
close-knit communities, especially friendships.3

Within this broad area I will pay close atten-
tion to eudaimonistic issues (especially, a virtue
ethics of personal character that focuses, as Aris-
totle does, if not to all of the conclusions that he
does, on human flourishing and on the centrality
of friendship to such flourishing).4 Given this nar-
rowing of the field, the question of an ethics of
memory can be restated: is it on occasion virtuous

or admirable to try to hold on to certain memories
and, at least sometimes, vicious and reprehensi-
ble to not try to forestall one’s forgetting?5 If, as
I shall argue, trying to hold on to certain memo-
ries is properly a source of ethical admiration, the
central reason is that keeping such memories is
one basic condition of human flourishing. Forget-
ting, especially in the context of intimate or thick
relations, may impoverish one’s life.

ii. an ethical reading of eternal sunshine and
the loss of memory

Montaigne expresses the ethical usefulness of
memory nicely: “Memory is an instrument of won-
drous service, without which judgment is hard put
to do its duty.”6 As Margalit notes, promise keep-
ing is an important example.7 One famous ver-
sion of this link between memory and judgment
is George Santayana’s remark, “Those who can-
not remember the past are condemned to repeat
it.”8 While Santayana’s comment is perhaps most
often associated with William L. Shirer’s use of
it as the moral epigraph for Rise and Fall of the
Third Reich, the general point about memory and
judgment is profoundly important for an ethics of
intimate relations. This point is also the driving
narrative concept of the witty screenplay by Char-
lie Kaufman and the delightful film directed by
Michel Gondry.9 The cost to good judgment re-
sulting from forgetting is on full display in Eternal
Sunshine of the Spotless Mind in two distinct but
intertwined dramas. First of all, Mary’s story (a
tragic if ultimately avenged subplot) provides an
example of that particular kind of forgetting that
increases the danger of the repetition of serious

The Journal of Aesthetics and Art Criticism 66:1 Winter 2008
c© 2008 The American Society for Aesthetics

78 The Journal of Aesthetics and Art Criticism

mistakes. Second, Joel’s story is a more general ex-
ample of the forgetful duplication of actions, the
disadvantages of which are importantly not lim-
ited to the repetition of mistakes.

In Eternal Sunshine Dr. Howard Mierzwiak di-
rects Lacuna Inc., a popular (especially at the
holidays) New York practice of memory erasure.
While his waiting room is filled with clients like
the woman seeking to be spared the painful mem-
ory of her dead dog, the booming business of the
outwardly kind and seemingly well-intentioned
doctor is the elimination of the memory of de-
structive love affairs for those clients who want
“to move on.” Mierzwiak asserts, “We provide
that possibility.”10 As Joel contemplates “erasing
Clementine”—who is in one way already lost to
him because she has just had her memory of Joel
erased—Dr. Mierzwiak remarks, “I suggest you at
least consider the potential pitfalls of a psyche for-
ever spinning its wheels.”11 Lacuna’s memory era-
sure business is exceptionally busy around Valen-
tine’s Day, during which time the film is set. The
doctor is aided by a cheerful young staff of libidi-
nous incompetents who mix the work of erasing
their clients’ now painful memories of past lovers
with some darkly farcical pleasure seeking of their
own. Perhaps the most vulnerable member of this
team is Mary, the young receptionist who is smit-
ten by the much older Mierzwiak and what she
initially sees as his deeply humanitarian practice.
Mary says, “To let people begin again. It’s beauti-
ful. You look at a baby and it’s so fresh, so clean, so
free. Adults [are] like this mess of anger and pho-
bias and sadness . . . hopelessness. And Howard
just makes it go away.”12

First, in Mary’s story the ethical cost of for-
getting serious mistakes is revealed to be the
thereby increased risk of repeating them and suf-
fering again the consequent harm to self and oth-
ers that follows. The audience learns late in the
film that Mary has already had a painful affair
with the married Dr. Mierzwiak.13 Ignorant of her
own past—having already had her bitter memories
of this affair erased—Mary flirts again with the
not-so-good doctor (who, it then becomes clear,
had encouraged Mary’s erasure process, though
his precise motives for doing so remain obscure).
Repetition of this disaster is averted only by the
last-moment arrival of Mierzwiak’s wife. The film
makes clear that Mary’s near repetition of her
tragic past results in part from the choice to erase
memories that left those motives leading to the

tragedy completely unchecked by regret. As Mon-
taigne observed, in the absence of memory, judg-
ment is prone to error.

Second, the film’s central tale is about the era-
sure of Joel’s memory of Clementine. While Joel’s
is the most broadly revealing example of the in-
strumental value of memory, the lessons of Joel’s
story are not focused on the risks of repeating
forgotten mistakes. The clever and complex way
this particular story is told puts it beyond simple
summary. It should suffice for present purposes
to note the following. The bulk of the narrative
concerns itself with an internal view of this era-
sure of Joel’s memory, which is cross-cut with an
external view of the same. Internally the narrative
proceeds, roughly, in reverse order—that is, we see
Joel’s most recent memories of Clementine erased
first and the history of their friendship unfolds be-
fore us, and him, in reverse. This “wiping” of Joel’s
memory proceeds relentlessly toward the last of
Joel’s recollections, that of his first meeting with
Clementine. This is further complicated by the
presumption of the film that, as Mierzwiak notes,
“there is an emotional core to each of our mem-
ories.”14 The narrative upshot of this is that the
exact reverse chronology of the story is less than
orderly, presumably a consequence of the emo-
tional dimensions of each memory—that is, their
relative conscious or unconscious importance to
Joel. Because midway into the process Joel begins
to resist and to try to hold on to certain powerful
memories, the erasure of his memory of his two-
year-long relationship with Clementine proceeds
imperfectly and, at times, hilariously.15 The master
narrative of the film into which this story is placed
finally leads Joel and Clementine to a jointly am-
nesic Valentine’s Day “reunion” on the beach at
Montauk. This unknowing reunion is aptly placed
at the start of the film so, like Joel and Clementine,
the audience does not know until looking back
from the very end that this scene is indeed a re-
connection rather than a first meeting.

This central story of Eternal Sunshine inge-
niously highlights how Joel’s forgetting of Clemen-
tine has various distinct but related disadvantages
that go well beyond his incurring the liability of
repeating past mistakes. First of all, as Joel be-
gins to have his memory of the bad times with
Clementine erased (those times which, in anger,
he is, at the outset of the procedure, glad to have
erased), he loses as well many good and tender
memories.16 Joel’s later change of heart and mind

Meyer Reflections on Comic Reconciliations 79

becomes clear when, as we come upon a soon-to-
be eliminated memory of an especially affection-
ate moment, he says to the technicians of Lacuna
Inc., “Please let me keep this memory; just this
one.”17 Clearly, the fresh start that is purchased
with a loss of memory results in the yet deeper loss
of the awareness of the honest fact that Joel did
once love Clementine, their obvious troubles (her
alcohol abuse as well as their intertwined prob-
lems with self-esteem) notwithstanding. In short,
if you come to forget the bad times, the good times
go too.

Second, if you come to forget the bad times—
times wherein you have been wronged or hurt—
you cannot remember enough to be able to for-
give. This point about memory and forgiveness is
central to the ethical reading of the film. More-
over, there is reason to think that this link between
memory and forgiveness (and thereby the ethi-
cal reading itself) is central to the film. This fun-
damental link is represented on two levels. First
of all, the powerful conclusion of the film begins
with a series of apologies and moments of some-
times unspoken and even uncertain forgiveness
between Joel and Clementine.18 This merciful end-
ing is made possible because, after their unknow-
ing “reunion” at Montauk, Joel and Clementine
receive from Mary Lacuna Inc.’s copies of their
own memory files. Indeed, this restoration of rec-
ollection is, for one, Mary’s way of seeking for-
giveness and making amends for her role in the
destruction of their memories.19

These files, complete with audiotapes, pro-
vide detailed evidence that Joel and Clemen-
tine have had memories of each other erased.
Joel’s and Clementine’s tapes—veritable litanies
of complaints about the other and the sad state of
their former friendship—are partially overheard
by each other toward the end of the film, which
has the initial consequence of rendering them an-
gry with each other and quite confused. How-
ever, the ultimate dramatic upshot of this influx
of mostly bitter memory brings both Joel and
Clementine, who are at this point only begin-
ning to absorb these recovered “quasi-memories,”
to a concluding cascade of apologies. In a scene
displaying their state of mutual confusion and
regret—especially about what the erasure of their
memory has brought about between them—Joel
says, “I’m sorry I yelled at you,” and Clementine
replies, “It’s ok.” She then quickly follows with,
“I really like you and I hate that I said mean

things about you.” Only moments later Joel says,
“I’m sorry, I’m so sorry.” Clementine follows, once
again, with “I’m sorry about all this.”20 This final
flood of contrition and forgiveness—shot in a gray
interior with a somber soundtrack—is the key that
allows the two main characters to move forth at
the very end of the film to begin to recover their
friendship. Yet, this move toward a powerful, if
also decidedly modest, comic ending is only the
first way that the film expressly connects memory
and mercy.

The link between memory and forgiveness em-
phasized in the final scene of the story is also set
up on a metaphorical level from the very outset of
the film. When Joel first hears Clementine’s name
he says, “It’s a pretty name. . . . It means merci-
ful, right? Clemency?”21 To fully understand this
link between Clementine and clemency, it is cru-
cial to begin by noting that for the bulk of the film
(the entire story of Joel’s erasure process, which
reveals most of what we know about their friend-
ship) it is strictly speaking not Clementine, but
Joel’s memory of Clementine, that is on-screen.22

Seeing Joel’s view of the erasure process, we real-
ize that he is not only stuck inside his mind, unable
to escape and stop the process, but we are also fully
in Joel’s world, seeing people and events as he and
only he remembers them. The strongly subjective
point of view of these memories is crucial to appre-
ciating the place of Clementine as the film’s sym-
bol of clemency, as well as Joel’s Muse of mercy.
For most of the film Clementine is simply a player
in Joel’s internal story, a figment of Joel’s imagina-
tion of their past friendship. In this particular nar-
rative locale, which echoes Joel’s portrayal from
the outset, Clementine personifies clemency. From
this viewpoint, to “erase Clementine” is not only to
erase access to memories that provide the oppor-
tunity for forgiveness (my point above about how
making use of this opportunity ultimately leads to
the comic ending of the film). Doing so is also to
erase the film’s main sign of forgiveness.

Furthermore, from the point of view of Joel’s
ethical character, to “erase Clementine” is to erase
one chief source supporting his disposition to for-
give. Joel’s life without memories of Clementine
is his life without connection to her friendly in-
spiration toward forgiveness, which would surely
impair Joel’s own talent for mercy. This under-
standing of Clementine as Joel’s Muse of mercy
is best seen within the context of Aristotle’s idea
of the friend as an “other-self.”23 (I will discuss

80 The Journal of Aesthetics and Art Criticism

why Aristotelian friendship is an especially apt
notion to invoke in this particular film below in
Section iii.) Right now the main point is this: in the
Aristotelian tradition (from Seneca to Montaigne
to Emerson) of the friend as an “other self,” the
friend is understood as a source of moral suste-
nance. For Aristotle, “[t]hose who wish for their
friends’ good for their friends’ sake are friends in
the truest sense.”24 In short, for one to forget a
friend who directly promotes one’s virtuous dispo-
sitions would be a form of ethical self-mutilation (a
point to which I will return in Section iv). This de-
struction undermines the seeking of individual and
shared progress toward happiness, understanding
“happiness” here as a kind of ethical flourishing
or eudaimonia. So “erasing Clementine” has the
poignant consequence for Joel of eliminating a
personally powerful prompting for his own ten-
dency to live out the virtue of mercy. The upshot
of this use of Clementine as Joel’s inspiration for
forgiveness is that, for the bulk of the story, he is
experiencing the erasure of that part of his mem-
ory that helps provide access to his disposition for
the virtue of clemency as well as related aspects of
eudaimonia that would arise from this source.

So “Joel’s Clementine” is a figure of forgiveness
and a Muse of mercy. Both points become quite
clear when at the end of the story both Clemen-
tine and Joel’s memory of Clementine are brought
back into the film and back into relationship with
him. As I outlined above, the result of this sym-
bolic return of memory combined with this reunit-
ing of the friends is not only that the possibility as
well as the spur to the disposition for mercy re-
turn but also that the penultimate scene, with its
repeated requesting and granting of forgiveness,
quite naturally leads to the powerful end of the
story. And it is this very progression that gener-
ates one final point in support of the ethical read-
ing of the film. Indeed this final point focuses di-
rectly on the moving and upbeat ending of the film.
The point is this: if you can’t remember the bad
times (and so, as argued above, you simply can’t
forgive), you also can’t achieve reconciliation. The
link between memory and reconciliation is the eth-
ical lesson of the comic conclusion of the film. If
we understand ‘forgiveness’ as it is often under-
stood (as a forswearing of attitudes like resent-
ment), forgiveness between friends may be only a
first step toward reconciliation.25 Within this story
reconciliation might be best understood as based
on a mutual forgiveness (one given and received in

both directions) and a further shared appreciation
of this merciful exchange. For one, such a recon-
ciliation is the final step toward reestablishing a
friendship of an Aristotelian sort.26 And reconcil-
iation is the unmistakable subject of the final scene
of the film.

Seeing this point is enhanced by the fact that
the end of the film has two distinct final scenes—
the moment of discussion in the hallway and
then the last images of Joel and Clementine play-
ing happily on the beach. Each scene exhibits a
separate aspect of reconciliation. First, consider
the ending dialogue of the film, which displays a
penetrating insight about renewed friendship. Af-
ter Clementine has left Joel’s apartment in anger
and confusion (in spite of their recent series of
apologies), Joel follows her into the hallway.

Joel (urgently): Wait.

Clementine (with slight exasperation): What?

Joel: I don’t know.

Clementine & Joel (impatiently talking over each other):
What do you want, Joel? I don’t know, just wait. I want
you to wait for just a while.

(Long pause broken by Clementine’s anguished exhale
that signals to Joel that he can approach her.)

Clementine: Okay.

Joel: Really?

Clementine (sternly): I’m not a concept, Joel. I’m just a
fucked-up girl who is looking for my own peace of mind.
I’m not perfect.

Joel (plaintively): I can’t see anything I don’t like about
you right now.

Clementine (forcefully): But you will. You will think of
things. And I’ll get bored with you and feel trapped be-
cause that’s what happens with me.

Joel (with an air of wise surrender): Okay.

Clementine (softening): Okay.

Joel (with a goofy smile): Okay.27

Clementine (whose anguish has now turned to laughter):
Okay.

The sheer affective power of this scene, a re-
sult of excellent acting, gives this insightful dia-
logue immense potency and authority. The friends
have found each other, once again. And with this

Meyer Reflections on Comic Reconciliations 81

rediscovery, and the shared awareness of the im-
perfect nature of their friendship, the comic con-
clusion has now been reached. The last ethical
connection has also been made: in the film, mem-
ory leads to acts of forgiveness that between these
friends lead to mutual acts of reconciliation.

Given Joel and Clementine’s lingering confu-
sion about all that has just transpired, the direc-
tor underlines this subtly upbeat closing dialogue
by cutting from sounds of Joel’s and Clementine’s
shared laughter in the hallway to a final shot of the
now-again-friendly couple playing on the beach
at Montauk. Whether this is a memory or a pro-
jected future or perhaps both we do not know.
For enhanced expressive power the exact middle
of the scene of reconciliation just set out has a
cue to a song, the lyrics of which underscore the
emotional roots of reconciliation.28 Over the fi-
nal shot of the reunited friends playing, we hear
these lyrics, sung at an unhurried tempo: “Change
your heart; look around you. Change your heart;
it will astound you. I need your loving like the
sunshine.” These lyrical reflections play over this
final shot of re-attained friendship—an ongoing
deeper, if also clearly imperfect, reunion—that has
been provided for by the friends’ acts of reconcilia-
tion. Indeed, taken together these two parts of the
final scene are a discerning example of a “remar-
riage” by the general standards of the comedies
of remarriage. This particular interpretation rep-
resents further ethical and aesthetic developments
that I will examine in the next section.

So far the ethical reading of the film goes as
follows. Memory is of indispensable instrumental
value to ethical judgment in various ways—all of
the following of which are on display in this film:
without painful memories, you are prone to repeat
past mistakes, resulting in further suffering; with-
out painful memories you also lose much in your
past that sustains you; without painful memories
you forget that which is necessary to be able to
forgive a friend; and without the forgiveness, the
opportunity for which is itself a gift of memory, an
act of friendly reconciliation remains elusive. As
Montaigne suggests, without memory judgment is
thwarted. And without good judgment there is lit-
tle or no opportunity for eudaimonia. Eternal Sun-
shine portrays the risks of unhappy, even tragic,
outcomes that result from lost memory. Such
risks in turn provide compelling reasons to try to
avoid forgetting even quite painful parts of one’s
past.

iii. comedies of remarriage, ethics, and memory

I have already called Eternal Sunshine a comedy—
I will also describe how it is more precisely a
twenty-first-century development of a comedy
of remarriage à la Stanley Cavell.29 Both claims
might well be questioned, and defending them
briefly will help us further understand how this
noteworthy film illuminates the role of ethics and
memory in remarriage comedies. A no less author-
itative source on the film than writer Charlie Kauf-
man says that he “realized early on that [Eternal
Sunshine] wasn’t a comedy.”30 Nonetheless, there
are well-known problems with discrediting an in-
terpretation of any film on the basis of a conflict-
ing understanding of the film from any one ma-
jor source (even an indispensable and justly pres-
tigious writer for a film, the success of which is
indisputably related to the story and the screen-
play).31 Furthermore, in the case of Eternal Sun-
shine we must remain aware, as I argued above, of
how director Michel Gondry realized Kaufman’s
screenplay, especially the comic conclusion.

I start with the ordinary notion of comedy as a
closing feature of a narrative. For present purposes
a comedy will be seen as a humorous story with a
happy ending or, more subtly, a humorous story
that concludes with a mere possibility of a happy
future for the main protagonists.32 The latter de-
scription of comedy fits some of Kaufman’s stated
hopes for Eternal Sunshine: “I never wanted to be
happy that they got together at the end. I didn’t
necessarily want it to be sad, but I wanted to leave
it up to the audience to decide.”33 Of course, the
film Gondry ultimately directed is, I have argued,
a comedy in the former sense—that is, it has a
straightforwardly happy ending. If the dialogue is
subtly happy, indicating only a bare possibility of
renewed friendship, the imagery and the lyrics of
the coda, as the reading in Section II shows, sug-
gest that the film is, in my sense, a simple comedy,
because the coda foreshadows some kind of happy
future for Joel and Clementine.

Yet this claim that Eternal Sunshine is best seen
as a straight-up comedy must be qualified in two
ways. First of all, I understand comic “happiness”
as some kind of eudaimonia. For present purposes,
a comedy—simple or subtle—reaches some con-
clusion with regard to the ethical well-being of
the main characters, especially as this is obtained
in part through their own efforts. In this sense
a simple happy ending is an ending portending

82 The Journal of Aesthetics and Art Criticism

genuine human flourishing for the main protag-
onists; a subtle happy ending allows only the pos-
sibility of such authentic flourishing. Second, such
happiness or flourishing might also be anxious or
troubled. This is a point to which I shall soon re-
turn, but the main idea is that happy endings might
well include some genuine anxiety, future human
happiness itself being a complex matter. Given
these two qualifications, “happily ever after” sug-
gests an ethical depth as well as an uncertain
future.

Eternal Sunshine is also a kind of romantic com-
edy, belonging in the final analysis to the gen-
eral family of remarriage comedies first identified
by Stanley Cavell. That this might be doubted is
obvious—for one, between Joel and Clementine
there is neither marriage nor divorce nor remar-
riage. In this respect Eternal Sunshine seems a
long way from a central exemplar of the remar-
riage comedy like The Philadelphia Story (George
Cukor, 1940). Starting to address this doubt, let me
briefly discuss some comments Cavell has made
regarding this subgenre. Cavell says that “mar-
riage is an allegory in these films of what philoso-
phers since Aristotle have thought about under
the title of friendship, what it is that gives value
to personal relations.”34 He notes, “Marriage in
these films may . . . be taken to stand for the idea
of friendship. . . . [T]he question of sealing or weav-
ing together the life of romance and of friendship,
while clearly taken in these films, almost without
exception . . . is rarely made explicit in the pair’s
conversations . . . though it should be seen as per-
vading them.”35 In remarriage comedies, as dis-
tinct from, say, Shakespearean comedies, friend-
ship supplies the only authentic basis for marriage.
Moreover, since in remarriage comedies “mar-
riage” stands first of all for some conception of
friendship, then the presence of literal marriage
and divorce is not a necessary condition for mem-
bership in the genre. One case in point here is the
metaphorical nature of the first marriage and di-
vorce in Frank Capra’s It Happened One Night
(1934) or the lack of either in Howard Hawks’s
Bringing Up Baby (1938).

In developing and then expanding Cavell’s basic
views I will focus on one main feature that clearly
is central to the remarriage comedy: “the drive of
its plot is not to get the central pair together, but
to get them back together, together again.”36 In
other words, what I take to be indispensable to
remarriage comedy is some kind of reconciliation

between erstwhile lovers, spouses, or friends.37 The
ethical reading so far suggests several points about
what I will call “remarriage reconciliations,” the
first group of which concerns comedies of remar-
riage in general; the others are essentially about
Eternal Sunshine.

First off, in these films a true marriage is not
reducible to either a ceremonial moment in time
or a legal institution. Instead such a marriage is,
or is grounded in, a standing willingness to take
part in a recurring series of remarriages. Put differ-
ently, commitment between friends—when deep
or especially trustworthy—is at times best seen as
a pattern of recommitment. This repetition sug-
gests something in general about reconciliation
between friends; a suitable “remarriage reconcili-
ation” is no one-time gesture. Such a reconciliation
is, for one, an ethical disposition—a propensity for
recommitment that, in order to have value, needs
to be lived over and over again. A full-fledged “re-
marriage reconciliation” is not a single act but a
targeted disposition to reconciliation. It is in this
way that such rapprochement is a crucial aspect of
eudaimonia. Against this background, what Eter-
nal Sunshine helps to further clarify is this: that one
key to the disposition to recommitment or remar-
riage or reconciliation is an acceptance, or even
better a shared acceptance, of risk and even ten-
sion between the friends. In Eternal Sunshine this
acceptance is signaled by Joel’s modestly cheerful
“Okay” at the end of the film. This is an act of ac-
ceptance that “living happily ever after” takes the
form of “living uncertainly ever after,” but doing
so in a spirit of memory, forgiveness, and reconcil-
iation.

Understood this way, the “remarriage reconcil-
iation” displays how it is that happy endings are
obviously just beginnings—more precisely, in this
genre, happy endings are “re-beginnings” or start-
overs that cry out for a memory of insult and in-
jury as well as forgiveness and friendly reconcilia-
tion. In classic remarriage comedies happy endings
are typically second chances for a more mature
couple. Such happy endings are also, in general,
the least stable and the least ethically satisfying
when they are not backed by a sturdy and dis-
cerning disposition for reconciliation. In the clas-
sic period of the Hollywood remarriage comedy,
one case in point is the madcap and ultimately
insecure remarriage at the end of Preston Stur-
gis’s Palm Beach Story (1942). Sturgis underlines
the insecurity of this particular remarriage by

Meyer Reflections on Comic Reconciliations 83

explicitly casting suspicion on the future of the
principal couple’s reconciliation by following the
ultimately sarcastic ending title “And they lived
happily ever after” with the skeptical closing cap-
tion “Or did they?” Possibly even more insecure
is the nervously fraught ending of the somber re-
marriage comedy His Girl Friday (1940), directed
by Howard Hawks. In short, reconciliation alone
(or perhaps in its limiting case in Hawks’s film
mere reconnection without forgiveness) within
this genre of films is no guarantee of a conven-
tionally successful future for the central couple,
let alone an ethically enlightened one.

It is also crucial to see that a “remarriage recon-
ciliation” is not a reconciliation at all costs—least
of all, as Eternal Sunshine crucially highlights, at
the cost of forgetting serious problems. That way
lies unhappiness, perhaps even tragedy. The ethi-
cally best happy endings in this genre are grounded
in reconciliations that are born of solid memory of
past troubles. Furthermore, as Eternal Sunshine
brings to the fore, ethically admirable comic con-
clusions can include a stalwart acceptance (and
best of all, a shared acceptance) of some sense
of risk and some anxiety about the future of the
friendship. Highlighting these insights in the main
theme of reconciliation within remarriage comedy
establishes for Eternal Sunshine a place among the
central exemplars of this rich style. In short, Eter-
nal Sunshine helps us better understand the risk
and anxiety of the “remarriage reconciliation,”
which is itself one key animating idea of this ethi-
cally powerful genre. Consider in this respect the
dialogue toward the end of The Philadelphia Story
where Tracy (Katherine Hepburn) says to Dexter
(Cary Grant) at the moment of their proposed re-
marriage: “Dext are you sure?” Dexter: “Not in
the least, but I’ll risk it, will you?” Tracy: “You
bet.”

Now, how does an appreciation of the genre of
comedies of remarriage help us deepen our un-
derstanding of the relation between ethics and
memory in Eternal Sunshine? First of all, “anx-
ious happy endings” are indeed a kind of classic, if
not always fully appreciated, conclusion for come-
dies of remarriage. This kind of anxious or unquiet
happy ending helps better explain the significance,
and the possible trajectories, of the comic conclu-
sion in Eternal Sunshine. Comedies of remarriage
emphasize how seemingly endless, but not mind-
less, repetition of commitment is the form that
commitment must often take to achieve the goal of

long-term, flourishing friendship. Given this em-
phasis, Joel’s insight at the end of the film has the
deepest ethical value when it is seen as a kind of
suitable “remarriage reconciliation.” The recon-
ciliation in the first part of the final scene has,
for one, less significance as an isolated act and
considerably more ethical value as evidence of a
discerning disposition to recommitment. The film
suggests that Joel wisely initiates acceptance of an
imperfect past and future with Clementine, and
also critically with himself as her partner (a point
to which I will return in Section iv). In doing so
Joel does signal a crucial understanding of the po-
tentially jarring nature of their friendship. But it
would be a mistake to assume that this single act of
acceptance (important though it is) represents all
of the ethical concerns presented by the anxious
happy ending of Eternal Sunshine.

If there must be a disposition for recommitment
within this, or most any, successful long-term ro-
mantic friendship, then Joel’s and Clementine’s re-
vealing series of “Okays” might also be taken to
signal their (beginning) allegiance to the deeper
insight that a disposition for reconciliation is a
cardinal virtue of true friendship. Alternatively,
they might not be understood to go beyond a
one-time assent to taking the calculated risk of
a flawed, or even a failed, future with each other.
Neither of these readings of this simple (but, since
also anxious, far from obvious or simple-minded)
comic ending seems required. Kaufman is right
about this point at least: the audience may de-
cide for itself if Joel and Clementine are likely to
persist as a romantic or even a friendly couple.
What is, however, beyond dispute (and thereby
not up to the audience to decide) is that if this,
or perhaps most any, romantic couple is to remain
within a flourishing friendship, they must live by a
key insight of the remarriage comedy—that when
commitment between friends and lovers takes its
most admirable and trustworthy form, it becomes
a discerning disposition to recommitment. This
still somewhat general form of “remarriage recon-
ciliation” is deepened when the remarriages be-
come sensitive to a further level of complexity:
over the course of an entire lifetime together the
friends must continually reconcile themselves to
their changing needs, both as individuals and as a
couple. Ultimately, it is in the service of this deeper
form of reconciliation that memory serves one of
its highest ethical callings. Memory can lead not
only to an act of reconciliation but also to a stable

84 The Journal of Aesthetics and Art Criticism

and discerning disposition to reconnect with one’s
closest friends in light of a fully remembered past
and an ever-evolving and often anxiety-provoking
complex of future needs.

In summary, the remarriage comedy—in both
classic and contemporary manifestations—is an
ethically significant genre at the heart of which
lies the “remarriage reconciliation.” This particu-
lar type of reconciliation presents a crucial point
of inquiry into some of the most significant and
complex relations between memory and an ethics
of intimate relations. It also provides an essential
opportunity for understanding the ethical and aes-
thetic aspects of what I have called “anxious happy
endings.” As of this writing, Eternal Sunshine is
arguably the most instructive and also the most
canonical twenty-first-century comedy of remar-
riage. Understanding Eternal Sunshine within the
context of the central concerns of the genre of re-
marriage comedy is crucial to understanding the
full ethical import of the anxious comic ending of
this admirable film. Moreover, Eternal Sunshine
clearly belongs in this intriguing genre because
it adds depth to our understanding of the central
idea of the “remarriage reconciliation,” including
its emphasis on self-aware and mutual acceptance
of personal and shared anxiety regarding the cou-
ple’s past and future friendship.

iv. the anxious happy ending, memory,
and personhood

As already outlined, memory is instrumental to
good judgment in part by helping enable desirable
outcomes, including the avoidance of repeated er-
ror and suffering as well as the realization of for-
giveness between persons in close-knit commu-
nity. It is also important, as mentioned in Mar-
galit’s discussion of ethics and memory, to see that
being a “person” can be seen as an achievement.38

On this view, being a (fuller or more admirable)
person is an accomplishment realized with the aid
of memory.39 Memory is essential to personhood
when the notion of a person is ethical—that is,
neither that of a bare human being nor a ratio-
nal agent. Instead, a person is an individual whose
ethical identity is constituted, in part, by his or her
close relations. In this sense a person’s achieve-
ment of personhood is dependent on some appro-
priation of past relations, which is in turn depen-
dent, in various ways, on memory. In the context of
an ethics of thick relations, to forget much, or even

some crucial parts, of one’s intimate relations is to
fail to fully realize—or if part of a downward trend
to begin to lose—one’s personhood. Consider the
deep loss when an aging parent does not recognize
one of his or her grown children. In short, memory
is essential to being a person in the first place. One
conspicuous reason for this is that memory is es-
sential for a person’s maintaining and developing
intimate relations.

I will explicate this by focusing on the ethical
value of what I have called the anxious happy end-
ing. A fuller appreciation of the anxious happy
ending has value for an individual (and his or her
friendships) because it is a critical vantage point
from which to view the state of his or her intimate
relations and ethical identity. In this regard an anx-
ious happy ending can be central to the achieve-
ment of his or her full personhood. When the pro-
cess of becoming a fuller person includes an anx-
ious happy ending to one episode of a life story,
this particular conclusion is less a final event than
an opportunity within an incomplete process. An
anxious happy ending is also an underappreciated
aesthetic notion which, given the present focus, is
grounded in a remarriage reconciliation.40 The two
main features of anxious happy endings are anxi-
ety and incompleteness, both of which are typically
connected with memory of past problems. Having
already indicated how anxiety and unease can ac-
company a “happy ending,” I will elaborate here
on the necessary incompleteness of the anxious
“happy ending.” Clearly enough, happy endings
can be, and often should be, viewed as beginnings.
Indeed, one is tempted to say that happy endings
typically should be viewed as beginnings because
the very happiness they promise depends on also
viewing the ending as a beginning.

The main advantage that results from seeing the
anxious happy ending as a beginning is the contrast
that this kind of unfinished ending provides with
the more static comic resolution. In a typical comic
resolution, a largely new and better society is cre-
ated or rediscovered in service of the complete
overcoming of an alternative unhappy social or-
der. In Frank Capra’s It’s a Wonderful Life (1946)
the vicious alternate world of “Pottersville,”
dominated by selfishness and brutish pleasure, is
overcome by the film’s final evocation of the co-
operative utopia of Bedford Falls wherein George
Bailey finds affirmation for his life. The point is not
that Capra’s film lacks a sense of uncertainty about
the future of the larger world; its very strength is

Meyer Reflections on Comic Reconciliations 85

that it is quite well attuned to worries about the
kind of society the audience might inhabit after
the end of World War II. But the attractions of the
utopian version of Bedford Falls are largely nostal-
gic, and nostalgia itself is one standard distortion
of memory, creating as it does a misleading ide-
alization of a seemingly familiar past. Moreover,
Capra’s film fails to use the uncertainty that gives
it some of its undeniable power in the service of a
more nuanced comic conclusion. The unquestion-
able popularity of such comic resolutions, wherein
the main problems of the story are solved or their
clear solution is reliably foretold, may be a direct
function of their static or settled nature.

Yet, static happy endings, especially corny
comic conclusions, invite little or no further re-
flection on the happy resolution of the story.41

Nonetheless, whatever legitimate satisfactions
such happy resolutions give rise to, they also tend
to be undercut by a lingering sense that the con-
clusion was either contrived or one-dimensional.
Roughly, if the conclusion is judged to be contrived
(say, via the use of a deus ex machina), then at some
level the plot failed to be convincing. If, however,
the conclusion is judged one-dimensional, then the
plot may have some convincing features, but it still
tends be undermined somewhat by its association
with elements of melodrama (for instance, when
the overt goals of the plot so emphatically drive a
particular character’s destiny that there is no seri-
ous chance for a strongly realized character to ac-
tually deepen the story). While static happy end-
ings are often a kind of straightjacket for comic
characters, anxious happy endings tend to ease
this predicament by allowing comic characters and
their stories to breathe with some of the reality and
uncertainty of the human condition.42 One main
problem with the static resolution is that, when
predominant, it tends to narrow the view of the
way endings function as beginnings.

The main aesthetic point here (that there are
two chief ways in which endings can function as
beginnings: static resolutions and anxious start-
overs) is made even sharper with the focus of the
remarriage comedy on re-beginnings. An empha-
sis on happy endings as more or less static resolu-
tions is radically opposed by the remarriage com-
edy’s underscoring of the continuity and the un-
certainty of change. The start-over-again style of
happy ending essential to the remarriage comedy
tends to be antithetical to any radical break with
the past or with our honest memory of it. The very

uncertainty of the future that results from this kind
of recollected continuity with past problems (and
the disavowal of a view of a future that is simply
resolved or otherwise fully harmonized) is central
to what the remarriage re-beginning emphasizes
within film comedy in general. The ethical read-
ing of Eternal Sunshine offers a case study on this
point, one upshot of which is this: just as the hu-
morous is more powerful when it is aptly paired
with the serious, the comic gains some aesthetic
import to the extent to which it approaches the
dilemmas that give life to tragedy.

At the end of Eternal Sunshine Joel’s reappro-
priation of his own ethical identity is a function
of his ability to realize some of the wisdom of
the anxious happy ending. At the point reached
by the precoda finale of Eternal Sunshine, Joel
does realize actual, if tentative, progress in achiev-
ing or recovering his ethical identity—indeed this
is one reason to view the coda as an image of
both past and future friendship. His personal eth-
ical progress is due to his start-over with his old
friend (who is, crucially, let us recall, an Aris-
totelian “other self” and thereby, for one, an in-
valuable reflection of Joel’s better self). This re-
newed friendship creates a higher and more trust-
worthy level of self-awareness for Joel, and it does
so here via an uncompromised memory of past
difficulty. Even if clarified in the light of reconcil-
iation, Joel’s and Clementine’s uncertain futures
are also overcast by a keen awareness of their
past problems. And, typical of an anxious happy
ending, no reliance is placed upon an unconvinc-
ing hope that all problems have been or can be
solved.

Nonetheless, Joel and Clementine’s anxious re-
marriage does display a quiet hope. This is so, in
part, because as a couple their ethical identity can
be, by the time of the comic conclusion, located
within their joint effort to start again without for-
getting, to start again in search of happiness amidst
insecurity. At the end of Eternal Sunshine Joel and
Clementine as separate persons also can be seen to
pursue individual identities that embrace anxiety
and uncertainty. Here starting over with the same
person is starting over both with the same friend
and with a self sharply aware of his or her own his-
tory. Such starting over is opposed to starting all
over again, as it were, from scratch; and it is one
admirable and trustworthy way to seek personal
flourishing. Such a restart foregrounds the role of
memory in the search for ethical identity.43 The

86 The Journal of Aesthetics and Art Criticism

anxious remarriage at the end of Eternal Sunshine
is admirable because it is joined with a wise accep-
tance of human limits, an acceptance that is in turn
made acute by an unsparing memory of a shared
past. In effect this ethical perspective applied to
the entire genre of remarriage comedies pays a
further aesthetic dividend. This wisdom of ground-
ing human aspiration in unstinting remembrance
of frailty and failure is perhaps one of the chief
claims to aesthetic eminence (often reserved by
some, like Aristotle and Hegel, for tragedy alone)
that comedy can make.

v. conclusion: the dignity of comedy and the
ethics of memory

This reading of Eternal Sunshine is evidence that
comedies of remarriage thrive in the twenty-first
century. Like the best classic remarriage comedies,
Eternal Sunshine also supplies the comic depth
necessary for ethical study—here a compelling
study in the ethics of memory and the aesthetics
of anxious happy endings. Margalit’s own reflec-
tions on ethics and memory suggest to him that
“[e]ven the project of remembering the gloomi-
est of memories is a hopeful project. It ultimately
rejects the pessimistic thought that all will be for-
gotten.”44 This ethical reading of Eternal Sunshine
adds something significant to Margalit’s decidedly
modest rejection of pessimism. Since holding on to
certain memories is an admirable trait and an as-
pect of human flourishing, it should be easier to
see how at their best the re-beginnings, the recon-
ciliations, and the remarriages that are essential
here to the anxious happy ending provide us an
encouraging vision of comic consolation. Indeed,
they present a kind of unromanticized yet astutely
hopeful comedy, because such a comedy strives
to live within the open-ended and memory-laden
boundaries of our less than ideal lives.45

MICHAEL J. MEYER
Department of Philosophy
Santa Clara University
Santa Clara, California 95053

internet: mmeyer@scu.edu

1. Avishai Margalit, The Ethics of Memory (Harvard
University Press, 2002).

2. Margalit, The Ethics of Memory, p. 7; Bernard

Williams, Ethics and the Limits of Philosophy (Cambridge
University Press, 1985), chaps. 8–9.

3. For an account that focuses on the morality (but not
the ethics) of memory, see Christopher Grau, “Eternal Sun-
shine of the Spotless Mind and the Morality of Memory,” The
Journal of Aesthetics and Art Criticism 64 (2006): 119–133.

4. Aristotle, Nicomachean Ethics, trans. Martin Ost-
wald (Indianapolis: Bobbs-Merrill, 1962).

5. I do not discuss here this question: are remember-
ing and forgetting properly subjects of ethical evaluation?
Basically, I think it would take my discussion too far afield
to take up this question in any detail. But I also endorse
Margalit’s discussion of this point (The Ethics of Memory,
pp. 56–58) wherein he notes that the uncontroversial prac-
tice of giving promises evaluative weight shows that mem-
ory is a proper subject of ethical evaluation. In short, tak-
ing promises seriously requires that memory be a subject
of normative evaluation. What I do add here to his account
of this general issue is the qualification “try to hold on to
certain memories.” I believe that such an effort is generally
presumed to be, at least prima facie, a proper source of nor-
mative evaluation, since there is much one can in fact do
to make such an effort (keeping pictures, writing journals,
collecting old letters, and the like). Note that it is crucial to
the plot of Eternal Sunshine that such journals and letters be
destroyed.

6. Montaigne, The Essays of Michel De Montaigne,
trans. M. A. Screech (London: Penguin Press, 1987), p. 738.

7. Margalit, The Ethics of Memory, pp. 56–58.
8. George Santayana, “Reason in Common Sense,” in

Life of Reason (New York: Scribners, 1905), p. 285.
9. Eternal Sunshine of the Spotless Mind (Focus Fea-

tures Universal Studios, 2004); Charlie Kaufman, Eternal
Sunshine of the Spotless Mind: The Shooting Script (New
York: Newmarket Press, 2004). Dialogue cited throughout
this article has been transcribed by the author directly from
the film and is only cited to page numbers in this text
when it does in fact coincide with The Shooting Script.
Unless otherwise noted, descriptions of intonation or in-
tention that parenthetically characterize cited dialogue are
my interpretations of the film and not director’s or writer’s
instructions.

10. Kaufman, The Shooting Script, p. 33. The character-
ization “destructive love affairs” comes from a clever faux
commercial for Lacuna Inc. (“Don’t forget, with Lacuna you
can forget”), which is an extra feature on the DVD.

11. Kaufman, The Shooting Script, p. 34.
12. Kaufman, The Shooting Script, p. 58.
13. A detail cut from the film but in The Shooting Script

(p. 107) is that the tragedy of Mary’s affair was to have been
further magnified by an abortion.

14. Kaufman, The Shooting Script, p. 38.
15. On my reading the film clearly encourages the au-

dience to root for Joel in this attempt to try to retain his
memories; the analysis that follows will elaborate this read-
ing.

16. “I’m erasing you and I’m happy,” Joel says early on
in the process; Kaufman, The Shooting Script, p. 44.

17. The filmmakers’ view of the importance of this par-
ticular scene (as well as some of how and why it was rewrit-
ten) is attested to by Charlie Kaufman, The Shooting Script,
p. 140.

18. Kaufman, The Shooting Script, pp. 125–129.

Meyer Reflections on Comic Reconciliations 87

19. What I call Mary’s “restoration of recollection” is,
to be more precise, symbolic. Joel and Clementine do not ac-
tually regain their memories as memories. All they are given
by Mary is evidence (the tapes) that they had certain past
experiences, evidence that they had memories of each other
and then had them erased. So, at most what they have access
to are a kind of “quasi-memories,” beliefs that may function
like some memories. In essence, the film asks the audience
to view all of this as a kind of metaphorical restoration of
memory.

20. Kaufman, The Shooting Script, pp. 125–127.
21. Kaufman, The Shooting Script, p. 9. Even though

Clementine’s response reveals at best ironic ambivalence
about her nominal relation to forgiveness (she responds in
anguish, “It hardly fits. I’m a vindictive little bitch, truth be
told” [The Shooting Script, p. 9]), she is revealed at the end of
the story to be more forgiving than her harsh view of herself
here suggests.

22. This fact about how to understand what the audi-
ence is presented during the erasure process is stated by Dr.
Mierzwiak himself. Kaufman, The Shooting Script, p. 62.

23. Aristotle, Nicomachean Ethics, p. 237.
24. Aristotle, Nicomachean Ethics, p. 219.
25. One standard account of forgiveness—the forswear-

ing of negative reactive attitudes like resentment—is found
in Jeffrie Murphy and Jean Hampton, Forgiveness and Mercy
(Cambridge University Press, 1988).

26. It should be said that this restoration of friendship
could also make it possible to end or downgrade a friendship
in a way that is not acrimonious but friendly. This is not to
conflate friendship and friendliness, even less to suggest that
one must restore a friendship in order to end it. It is simply to
note that, between former friends, a friendly end to relations
typically shows a flicker of the former friendship in addition
to the general virtue of kindness.

27. Joel’s smile is as important, and in a way as affect-
ing, as Clementine’s laughter. On the importance of the lucid
smile of self-awareness (and the relation between this and
laughter) see Simon Critchley, On Humor (Thinking in Ac-
tion) (London: Routledge, 2002), pp. 107–111.

28. The song is “Everybody’s Got to Learn Sometime”
by James Warren (EMI, 2005).

29. Stanley Cavell, Pursuits of Happiness: The Holly-
wood Comedy of Remarriage (Harvard University Press,
1981).

30. Kaufman, The Shooting Script, p. 136.
31. While a film is, in this respect, more complex than

a single-author literary work, an excellent recent discussion
of this complex topic is Steven Davies, “Author’s Intentions,
Literary Interpretation, and Literary Value,” The British
Journal of Aesthetics 46 (2006): 223–247. The strategy of the
present analysis is, in Davies’s terms, a “value-maximizing”
one with an additional focus, as outlined above, on specifi-
cally ethical concerns.

32. The place of humor within the genre of comedy is
a very complex matter. Since my analysis does not turn on
uncovering much of this complexity, I will leave this topic
unexamined except to basically endorse the view of humor
within comedy that Gerald Mast called the “comic climate.”
See Gerald Mast, The Comic Mind: Comedy and the Movies,
2nd ed. (University of Chicago Press, 1979), especially pp.
9–13.

33. Kaufman, The Shooting Script, p. 143.
34. Stanley Cavell, Cities of Words: Pedagogical Letters

on a Register of the Moral Life (Harvard University Press,
2004), p. 5.

35. Cavell, Cities of Words, p. 39.
36. Cavell, Pursuits of Happiness, p. 2 (emphasis in orig-

inal).
37. Other key features include that “upon separation

the woman tries a regressive tack, usually that of accepting as
a husband a simpler [man]” (Cavell, Pursuits of Happiness, p.
31). In this connection consider Clementine’s relations with
the adolescent figure of Patrick. Second, “a running quarrel
is forcing apart a pair who recognize themselves as having
known one another forever” (ibid.). In this connection con-
sider the dream sequence where Joel and Clementine, or at
least Joel’s memory of her, return to Joel’s childhood. Finally,
for the closing locale of a “green world” as a symbolic spa-
tial location for ethical transformation, consider the scene
on the beach at Montauk in the coda.

38. Margalit, The Ethics of Memory, p. 46.
39. There is an obvious ambiguity here between the

achievement of personhood, understood as a threshold no-
tion, and the achievement of one’s highest possible degree
of personhood understood not as a threshold but as an as-
cending matter of degree. As the difference between these
two notions of the achievement of personhood is not cen-
tral to the larger matter at hand, I will not unpack it further
here except to note the prima facie ethical admirability of
ever-higher degrees of personhood.

40. Anxious happy endings are decidedly not limited
to remarriage comedies. In film comedy one of the most
powerful and enduring is Charlie Chaplin’s justly famous
ending for City Lights (Charles Chaplin, 1931).

41. The sobriquet “capracorn” does seem to fit Capra’s
tendency toward more static comic resolutions, especially
Mr. Deeds Goes to Town (1936), Mr. Smith Goes to Wash-
ington (1939), and Meet John Doe (1941).

42. Two points of clarification are perhaps worth men-
tioning here. First of all, I have introduced a pair of basic
distinctions: subtle versus simple happy endings as well as
anxious versus static happy endings. These can be seen to
cut across each other, as one might have either an anxious
subtle happy ending or an anxious simple happy ending. I
think Eternal Sunshine as it was shot (with the coda) is of
the latter type, but it would have been of the former type
without the coda. Second, another manifestation of the dif-
ference between the static comic resolution and the anxious
happy ending exhibits itself as a general tendency to read
any ambiguous ending in one or another of these two basic
ways.

43. This point can be understood on several (in-
creasingly complex) levels because both the self and the
friend are somewhat changed by the action of the story.
On this point see Cavell’s discussion of The Lady Eve
(Preston Sturges, 1941) in his Pursuits of Happiness, pp.
47–70.

44. Margalit, The Ethics of Memory, p. 82.
45. I would like to thank Robert Lovering, Robert

Audi, Mark Ravizza, Daniel Ostrov, Larry Nelson, Lori
Zink, and an anonymous reviewer for this journal for help-
ful suggestions and comments on earlier drafts of this
article.

CHRISTOPHER GRAU

Eternal Sunshine of the Spotless Mind
and the Morality of Memory

The film Eternal Sunshine of the Spotless Mind
(Michel Gondry, 2004) is one of those movies
that people tend to either love or hate.1 Critics
generally raved about it, but if you look on web-
sites that allow people to post their own reviews,
you find a fair number of “one-star” ratings and
complaints that the film was confusing, pre-
tentious, or just plain boring. On the other hand,
those who like the film tend to really like it,
giving it five stars and admitting to having seen
the film multiple times in the theater. Why do
the fans of this film seem so, well, fanatic in
their devotion? Although I think much of their
appreciation has its base in the sensitive and cre-
ative direction of Michel Gondry, the clever
script from Charlie Kaufman, the beautifully
melancholy score by Jon Brion, and the impres-
sive performances by all the actors involved,
I also think it is not crazy to suggest that the
philosophy of the film helped it to achieve the
cult-like status it now enjoys.2

What, exactly, do I mean by saying that this
film has a philosophy? Well, I don’t just mean
that it explores philosophical ideas. It does this
very effectively, but it also offers something
more: in the course of exploring these ideas, it
implicitly offers a philosophical position. That
is, it does not just raise certain deep questions, it
suggests answers to those questions. Since it is
a movie and not a journal article, the position
that is gestured at does not come to us by way of
an explicit argument, but it is one that I think
can be unpacked and defended. Accordingly,
here I will be attempting to make explicit the
philosophical perspective that I take to be
implicit in this original and moving film.3

I. FORGET ME NOT

Eternal Sunshine of the Spotless Mind (Eternal
Sunshine) is a story about a group of people
who have access to a peculiar and powerful
technology. Thanks to the work of one Dr.
Mierzwiak (Tom Wilkinson) and his company
Lacuna, Inc., several characters are able to
undergo a process by which they have the mem-
ories of other people erased from their minds in
order to lessen the suffering that these painful
memories can cause. After watching the film, it
is hard not to dwell on the possibility offered to
the characters of Joel (Jim Carrey), Clementine
(Kate Winslet), and Mary (Kirsten Dunst). If
you could choose to erase someone from your
life, would you? Even if you personally would
not choose to undergo such a procedure, do you
think someone else should have that sort
of choice open to him or her? If the memories
of a particular incident or relationship are truly
causing someone tremendous pain, shouldn’t
they have the option of removing those memo-
ries, provided that it can be done safely and
effectively?

The film is wonderfully nuanced and subtle,
and thus not surprisingly it doesn’t offer us easy
or obvious answers to these sorts of questions.
Nonetheless, the general sense one gets from
the film is that the memory-removal technology
exhibited in the movie does not, in fact, allow
for the “eternal sunshine” referenced in the title.
Indeed, it is hard to imagine someone leaving
the theater thinking that pursuing such a tech-
nology would be a good thing.4 Why not? Well,
the film shows us some rather unfortunate

120 Thinking Through Cinema: Film as Philosophy

consequences that result from the use of the
technology: Mary, Joel, and Clementine, as well
as others connected to them, all experience pain
and heartbreak as a result of the supposedly
secret procedure going awry through various
leaks. However, what is more important for my
purposes is the fact that the film seems to
suggest that the memory-removal technology is
problematic even if the glitches and leaks could
be worked out. There is a sense of tragedy in
Joel’s realization (while in the middle of the
procedure) that he does not want to lose his
memories of Clem, and the sadness the viewers
feel with him is not lifted by the thought that he
will eventually be ignorant of the loss. On the
contrary, awareness of the future ignorance
seems to compound the sadness: that he will
soon be clueless is no cause for celebration. The
harm done by this procedure does not seem to
be fully accountable in terms of the harm the
characters consciously feel. In going through
the philosophical issues that are raised by the
film, I hope to offer an account of why the sense
of tragic loss suggested by the film resonates
with viewers, and why the implicit philosophi-
cal position assumed by the film is a respectable
and defensible one, even if it can at first appear
to be quite puzzling and controversial.5

II. UTILITARIANISM

In some ways the most obvious and sensible
response that could be made to the question “Is
the use of such memory removal technology a
good thing?” is what philosophers would call a
traditional utilitarian response. Traditional or

classical utilitarians (such as Jeremy Bentham,
John Stuart Mill, or Henry Sidgwick) thought
that the right action is the one that brings about
the most happiness overall, where happiness is
understood in terms of pleasure and the avoid-
ance of pain. When deciding on what to do, the
utilitarian does his or her best to calculate the
possible consequences of the choices that lay
before him or her. The morally right act is the
one that (among the possible actions open to the
person) will result in the most happiness and
the least suffering, and so the utilitarian will
always strive to choose those actions that are
most likely to increase overall happiness and
minimize overall suffering. Accordingly, if a
memory-removal procedure can function in
such a way that it brings about more happiness
than would otherwise be possible, the use of
such a procedure is not only justified, but in fact
morally required on utilitarian grounds.

Now it should be pointed out that there is a
big “if” in the claim above—it is not at all clear
whether this sort of procedure could be imple-
mented in such a way that it would increase
happiness overall. In Eternal Sunshine the pro-
cedure seems far from foolproof. Indeed, we see
fools implementing it (a stoned Stan (Mark
Ruffalo) and his dimwitted sidekick Patrick
(Elijah Wood)) and they do a thoroughly medi-
ocre job.6 We also see that the acquaintances of
Clementine fail to keep her procedure a secret
and in the process cause Joel no small amount
of misery. In addition, the memory-removal
procedure that Mary undergoes seems to
increase rather than minimize the pain and
suffering for everyone affected by her affair
with Mierzwiak. These and other considerations
would lead many people to conclude that the
procedure as displayed in the film does not tend
to maximize happiness overall.

The question remains, however, whether
such a process could be streamlined so as to
reliably minimize the suffering of those under-
going the procedure while not causing signi-
ficant harm to anyone else. Putting aside the
glitches and complications present in the film, it
is natural to wonder: If memory removal was
reliable, efficient, safe, and effective, are there
still reasons to reject it?

One might plausibly argue that painful mem-
ories stay with us for good reason: they allow us
to learn valuable lessons from the past and thus

FIGURE 1. Joel (Jim Carrey) undergoes a preliminary brain
scan in order to create a “memory map” that will be used to
erase his memories of Clementine (Kate Winslet).

Grau Eternal Sunshine of the Spotless Mind and the Morality of Memory 121

be better prepared for the future. This is no
doubt often the case, and in a situation in which
it appears that the removal of memory would
limit the person in this way (by denying him or
her useful information), such a procedure would
probably not be for the best (and thus not
“maximize utility”). However, there are cases in
which painful memories seem to do much more
harm than good, and where any lessons that
could be derived from the memories could
presumably be learned via other routes. In
those kinds of cases, it seems that the misery
avoided by memory removal would more than
counterbalance any possible benefits that would
normally arise from retaining the memories. It
seems, then, that the utilitarian response to
whether such a procedure is justified should be
a cautious and conditional “yes”: if suffering
can be minimized in a particular case, then such
a procedure is appropriate in that case. In
circumstances in which the use of memory
removal would increase overall happiness, the
use of such a procedure is, on utilitarian
grounds, a morally good thing. Moreover, as
I suggested earlier, utilitarianism would seem to
require the use of such a procedure if it was the
most efficient means of maximizing utility.7 For
the utilitarian, the goodness or badness of memory
removal hinges solely on the consequences, and
if we can ensure that those consequences are
beneficial overall, such technology would be
something to welcome rather than reject.

III. THE EXPERIENCE MACHINE SHIFTS INTO REVERSE

Many people will feel that the approach we have
been considering, though intuitive in many
ways, is somehow too crude. The worry is that
even if the procedure can reliably maximize hap-
piness overall (and minimize suffering) there is
still something wrong with it. Memory removal
seems problematic in a way that cannot fully be
made out within the utilitarian framework—a
loss has occurred even though we cannot expli-
cate the loss in terms of lost utility or happiness.

We can get at one reason why the procedure
in Eternal Sunshine seems so troubling by con-
sidering a classic example that is often used to
raise doubts about the hedonistic assumptions
that lie behind traditional utilitarianism. In his
1971 book Anarchy, State, and Utopia, Robert

Nozick introduced a thought experiment that
has become a staple of introductory philosophy
classes everywhere. It is known as “the experi-
ence machine.”

Suppose there were an experience machine that
would give you any experience you desired. Super-
duper neuropsychologists could stimulate your brain
so that you would think and feel you were writing a
great novel, or making a friend, or reading an inter-
esting book. All the time you would be floating in a
tank, with electrodes attached to your brain. Should
you plug into this machine for life, preprogramming
your life’s desires? … Of course, while in the tank
you won’t know that you’re there; you’ll think it’s all
actually happening. Others can also plug in to have
the experiences they want, so there’s no need to stay
unplugged to serve them. (Ignore problems such as
who will service the machines if everyone plugs in.)
Would you plug in? What else can matter to us, other
than how our lives feel from the inside?8

Nozick goes on to argue that other things do
matter to us: for instance, that we actually do
certain things, as opposed to simply have the
experience of doing them. Also, he points out
that we value being (and becoming) certain
kinds of people. I do not just want to have the
experience of being a decent person, I want to
actually be a decent person. Finally, Nozick
argues that we value contact with reality in
itself, independent of any benefits such contact
may bring through pleasant experience: we
want to know we are experiencing the real
thing. In sum, Nozick thinks that it matters to
most of us, often in a rather deep way, that we
be the authors of our lives and that our lives
involve interacting with the world, and he
thinks that the fact that most people would not
choose to enter into such an experience machine
demonstrates that they do value these other
things. As he puts it: “We learn that something
matters to us in addition to experience by imag-
ining an experience machine and then realizing
that we would not use it.”9

One way to think about the procedure pre-
sented in Eternal Sunshine is to consider it a
kind of reverse experience machine: rather than
give you the experience of your choice, it
allows you to take away experiences that you
have retained in your memory. Similar philo-
sophical issues arise, as the worry is that in both

122 Thinking Through Cinema: Film as Philosophy

cases we are achieving pleasure (or the avoid-
ance of pain) at the cost of truth. Elsewhere,
I have discussed Nozick’s thought experiment
in the context of the character Cypher’s (Joe
Pantoliano) choice in the film The Matrix (The
Wachowski Brothers, 1999).10 There, I argued
that our natural aversion to sacrificing know-
ledge of the truth for happiness can be under-
stood as the expression of some of our most
basic values, and that these values are perfectly
legitimate and need not be threatened by a
hedonistic outlook that claims that only pleasur-
able conscious experience can ultimately have
value in itself.

Not surprisingly, I think something similar
can be said about the memory-removal procedure
offered in Eternal Sunshine. Even if the use of
such a procedure would maximize happiness, it
is understandable and justifiable for someone to
refuse such a procedure on the grounds that they
do not want to “live a lie.” To think otherwise is
to forget that many of us value the truth in a
way that cannot simply be explained in terms of
the pleasure that knowledge of the truth often
brings or makes possible. Our reluctance to
endorse (or undergo) a memory-removal pro-
cedure is one expression of this basic value we
place on the truth for its own sake.

Toward the end of Eternal Sunshine, Mary
finds out that she has undergone the memory-
removal procedure and decides that what
Mierzwiak has done is horribly wrong. This
realization prompts her to return the medical
files of all his previous patients, telling them
that she has done this to “correct” the situation.
In the shooting script for the film there is an
additional bit of dialogue that further suggests
that her actions are motivated by considerations
similar to the sort we have been considering.

MARY: Patrick Henry said, “For my part, whatever
anguish of spirit it may cost, I am willing to know the
whole truth; to know the worst, and to provide for it.”
I found that quote last night. Patrick Henry was a
great patriot, Howard.11

Unfortunately, those not inclined to share this
intuition with Mary, Patrick Henry, or Nozick
(that truth has value that is independent of the
good consequences knowledge of the truth can
bring) are likely to complain that this position
stands in desperate need of justification. Why is

the truth valuable in itself? Why should we think
it good to know the truth in situations in which it
brings only misery? The natural response (we
just do value the truth in this fundamental and
basic way) is not likely to sway the person who
thinks a memory-removal procedure is unprob-
lematic. Although everyone agrees that justifica-
tions have to come to an end somewhere, rarely
do philosophers agree just where a proper end-
ing resides. The common response to Ludwig
Wittgenstein’s famous claim “my spade is
turned” (that is, I have hit bedrock—I have
exhausted justifications) is to tell him to pick up
the damn spade and keep digging!12 I am not
sure that much can be said to resolve this sort of
dispute, but one point that can be made is to
remind the opponent that he or she, too, hits
bedrock eventually, that is, a point at which he
or she can no longer provide a justification for
his or her own valuation. To the question, “and
what justifies the value you place on pleasant
conscious experience?” it seems little can be
said. This sort of concern appears to be some-
how self-justifying or beyond justification. If
this is right, it is unclear why we should not
allow that other concerns might well be simi-
larly foundational or beyond justification.

A further justification for valuing the truth
may not be possible; however, it is possible to
say a bit more by way of explanation regarding
why many people hold this value. Colin
McGinn has described the threat of general
epistemological skepticism as tantamount to an
individual discovering he or she is in a kind of
“metaphysical solitary confinement.” If we do
not know what we think we know, then we are
in effect cut off from the world. If the skeptic is
right, it turns out that our mind does not have
the kind of interaction and relationship with
reality that we ordinarily take it to have, and
this possibility is understandably disturbing to
us. As McGinn puts it, we want our mind to be a
window onto the world, not a prison.13

Merely losing a portion of one’s memory is
certainly not equivalent to the sort of radical
ignorance that epistemological skeptics enter-
tain, but it does involve a related variety of
detachment from the world. Having undergone
a memory-removal procedure, the individual
has consented to, if not a metaphysical prison,
then at least a pair of metaphysical blinkers and,
worse yet, he or she has consented to make

Grau Eternal Sunshine of the Spotless Mind and the Morality of Memory 123

himself or herself ignorant of that very choice.
The individual chooses to cut himself or herself
off from the world—his or her mind represents
the world less accurately than it did before and,
accordingly, he or she is slightly closer to the
isolation and solipsism that make skepticism
threatening.14 We have a very natural desire not
to be cut off from the world in this way, and
thus it is not surprising that the removal of
memories disturbs us in a manner that cannot
simply be cashed out in terms of future unhap-
piness. The fact that in Eternal Sunshine the
memory removal involves isolating a person
from someone who was previously very close
makes the use of the procedure all the more
disturbing: it is not just a metaphysical relation-
ship that has been severed, but a personal and
emotional one.

IV. WHAT YOU DON’T KNOW CAN HURT YOU

Granting that voluntary removal of one’s mem-
ories seems to clash with the value that many
people place on knowing the truth about them-
selves and the world, a further related question
arises: Does the sort of memory loss exempli-
fied in Eternal Sunshine involve an actual harm
or misfortune to the person who undergoes the
procedure? It is quite natural for people to think
initially that such a procedure cannot be said to
harm the person if it produces no unpleasant
effects for the person. (How can I be harmed if I
do not consciously experience the harm?)
Although this seems straightforward enough, on
reflection we can see that it is far from obvious
that this simple notion of harm will suffice.
Consider Thomas Nagel’s comments on the
view that harm must necessarily be experi-
enced.

It means that even if a man is betrayed by his friends,
ridiculed behind his back, and despised by people
who treat him politely to his face, none of it can be
counted as a misfortune for him so long as he does
not suffer as a result. It means that a man is not
injured if his wishes are ignored by the executor of
his will, or if, after his death, the belief becomes
current that all the literary works on which his fame
rests were really written by his brother, who died in
Mexico at the age of 28.15

Nagel reminds us that many situations that we
would naturally want to characterize as invol-
ving harms would have to be redescribed if we
want to embrace the narrow view that harms
must be experienced. Elaborating on Nagel’s
insights, Steven Luper helpfully distinguishes
between what he calls “harms that wound”
versus “harm that deprive.”16 We can under-
stand the harms that Nagel speaks of as harms
that may not wound but do deprive the person
of some good, and both Luper and Nagel suggest
that a sensible account of harm should be able to
incorporate these latter types of misfortune.17

If this approach is correct, then it would seem
that the deprivation of the truth that Joel,
Clementine, and Mary undergo in Eternal Sun-
shine could rightly be seen as a form of harm or
misfortune.18 The fact that it is something they
bring on themselves does not change this, for
we allow that people often (knowingly and
unknowingly) harm themselves in other ways.
(The film in fact implicitly supports this notion
through characterizing Clementine as self-
destructive, Mary as easily manipulated, and
Joel as a depressive—just the types of people
who could and would harm themselves.) The
harm here is not as dramatic or obvious as some
other forms of self-abuse, but it is nevertheless
genuine: they have sacrificed a part of their
minds and in the process blinded themselves to
a part of the world.

V. IMMANUEL KANT ON DUTIES TO ONESELF

So far I have suggested that memory removal is
morally problematic because it involves a clash
between fundamental values: our concern with
knowing the truth comes into tension with our
desire for happiness. Undergoing such a proce-
dure inevitably involves sacrificing the concern
for truth and, accordingly, we are inclined to see
the person who has undergone such a procedure
as having been harmed through deprivation of
the truth. Is there more to say regarding the sort
of harm that one undergoes here? I think there
is, and I think we can get at a deeper appreci-
ation of the harm to self that memory removal
involves through a consideration of some ideas
from Immanuel Kant.

Kant famously proclaimed that persons are
unique: everything else in the world is a thing

124 Thinking Through Cinema: Film as Philosophy

and thus has a price, but persons alone deserve a
kind of treatment that involves recognizing their
value as beyond price. Persons, because of their
capacity for freedom and rational agency, have
a dignity that is incommensurable and priceless.
Accordingly, persons deserve respect. Kant
thought that we needed to be consistent in our
thinking on these matters, and that means we
have to acknowledge that you have a duty to
treat yourself with respect and never to use
yourself solely as a means to an end. Accord-
ingly, he argued that morality prohibits both
suicide and many forms of self-mutilation. In
the Groundwork, he succinctly lays out his
reasons for this view.

First, as regards the concept of necessary duty to
oneself, the man who contemplates suicide will ask
himself whether his action can be consistent with the
idea of humanity as an end in itself. If he destroys
himself in order to escape from a difficult situation,
he is making use of his person merely as a means so
as to maintain a tolerable condition till the end of life.
Man, however, is not a thing, and hence is not some-
thing to be used merely as a means; he must in all his
actions always be regarded as an end in himself.
Therefore, I cannot dispose of man in my own person
by mutilating, damaging, or killing him. (It belongs
to ethics proper to define this principle more
precisely, so as to avoid all misunderstanding, e.g., as
to the amputation of the limbs in order to preserve
myself, as to exposing my life to danger with a view
to preserve it, etc. This question is therefore omitted
here.)19

Many have mocked Kant’s remarks, wonder-
ing if his prohibition should include such hor-
rific acts as ear piercing or haircuts.
Acknowledging that we may want to make
allowances for the permissibility of suicide and
bodily mutilation under certain circumstances,
we can still agree with the spirit of Kant’s
claims here: there is something disturbing about
the idea of self-manipulation that parallels the
disturbing aspects of manipulating others, and
consistency suggests that we should recognize
that cases of treating oneself solely as a means
are morally problematic for the same reasons
that objectifying others is wrong. Just as it is
wrong to use others for advantage (even their
own advantage) in ways that do not recognize
their humanity, it is wrong to objectify oneself

simply for the sake of some supposed advantage.
As Kant says elsewhere: “Self-regarding duties,
however, are independent of all advantage, and
pertain only to the worth of being human.”20

It is a natural extension of Kant’s view to
criticize the process we see in Eternal Sunshine
on the grounds that it involves a type of morally
problematic self-objectification. Part of what is
so disturbing about the memory-removal proce-
dure is that it is in fact a form of self-mutilation:
in order to “maintain a tolerable condition” one
uses oneself as a mere means and thus manipu-
lates oneself as though one were an object
rather than a person deserving of respect.
Indeed, the kind of manipulation involved here
is more obviously problematic than the sort of
bodily mutilation Kant mentions. After all, what
is mutilated in this case is not merely one’s body
but one’s mind, and thus the violation of one’s
rational nature is frightfully direct. Memory
removal bears closer similarities to the sort of
mind manipulation that Kant had in mind when
he rejected the idea of rehabilitating prisoners.
James Rachels, summarizing Kant’s view,
explains the rationale behind Kant’s opposition
to rehabilitation.

[T]he aim of “rehabilitation,” although it sounds
noble enough, is actually no more than the attempt to
mold people into what we think they ought to be. As
such, it is a violation of their rights as autonomous
beings to decide for themselves what sort of people
they will be. We do have the right to respond to their
wickedness by “paying them back” for it, but we do
not have the right to violate their integrity by trying
to manipulate their personalities.21

Kant’s view is obviously controversial, but it is
easy enough to understand his concern, at least
when considering certain types of rehabilitation.
Take the film A Clockwork Orange (Stanley
Kubrick, 1971): in it, a young thug named Alex
(Malcolm McDowell) is captured and undergoes
“aversion therapy” that makes him unable to
commit violent acts but does nothing to remove
his immoral desires or convince him of the
wrongness of what he has done. He becomes
mechanical, like clockwork, rather than a free,
rational agent. It is precisely the sense that Alex
has been unjustly manipulated that causes us to
have sympathy for an otherwise vile person.
Even if he is a criminal who has committed

Grau Eternal Sunshine of the Spotless Mind and the Morality of Memory 125

countless immoral acts, that does not give
society the right to treat him as though he is
merely a broken mechanism rather than a person.
Manipulating someone’s mind is a particularly
robust and offensive way to fail to grant him or
her the respect that all people deserve.22

One might think the parallel between mani-
pulative rehabilitation and self-induced memory
removal fails because the case of memory
removal involves a person voluntarily consent-
ing to the manipulation while the criminal does
not (presumably) consent to the rehabilitation.
This brings up some rather thorny issues regard-
ing the role of consent vis-à-vis Kantian ethics.
I am inclined to think that Kant’s account has to
involve more than simply consent in order for
an act to show proper respect, but for our pur-
poses here we can leave this debate aside, for it
seems quite likely that the person post memory
removal is likely not to consent to the procedure
that has been performed on him or her even if
he or she did consent prior to removal. (Mary
exhibits this pattern rather clearly in Eternal
Sunshine.) The postprocedure person falls quite
squarely into the class of persons who have had
their integrity and personhood violated through
the kind of manipulation that Kant criticized.

The way the memory-removal procedure
creates a later self that may not approve of the
earlier self’s choices brings to mind another
parallel, one that the film highlights in a partic-
ularly vivid fashion. The sadness we feel for
both Clementine and Joel parallels the sort of
sadness felt for people who, out of misery and
desperation, start down a path of self-obliteration
through drugs or alcohol. It is no coincidence
that Clementine is characterized as an alcoholic,
nor that Joel often appears so depressed as to be
borderline suicidal. Their choice to utilize the
memory-removal technology is presented as
being of a piece with their other self-destructive
tendencies. Kant would presumably agree that
these behaviors all involve a morally problem-
atic form of self-destruction. Discussing alcohol
(and suicide), he remarks:

For example, if I have drunk too much today, I am
incapable of making use of my freedom and my
powers; or if I do away with myself, I likewise
deprive myself of the ability to use [my powers].
So this conflicts with the greatest use of freedom, that
it abolishes itself, and all use of it, as the highest prin-

cipium of life. Only under certain conditions can
freedom be consistent with itself; otherwise it comes
into collision with itself.23

The removal of memories can be plausibly seen
as a limitation on one’s freedom, just as Kant
suggests both drunkenness and suicide limit
freedom. (The cliché “knowledge is power”
rings true here: the self-imposed ignorance
brought on through memory removal limits
your power and your freedom through limiting
your options.) As with the other cases that Kant
discusses, utilizing one’s freedom in order to
remove one’s memories involves a kind of
contradiction: you attempt to use your freedom
in order to limit your freedom. On Kant’s
approach, we have no right to do this to
ourselves, regardless of the convenience or
advantage of such a procedure.

I do not want to suggest that Kant’s positions
on suicide, self-mutilation, or rehabilitation are
clearly correct or uncontroversial—they are not,
and many smart and able philosophers have
criticized them. What I do want to claim is that
his overall position and the way it manifests
itself in these particular cases is both insightful
and worthy of consideration, and that the
insights Kant offers us apply rather nicely to the
topic at hand, that is, the ethics of memory
removal. Kant offers a rationale for why harm-
ing oneself in certain ways is particularly dis-
turbing and morally problematic. In cases of
suicide, self-abuse, and (I have argued) memory
removal, we see agents treating themselves
solely as a means to an end rather than as ends
in themselves. There is a failure of self-respect,
and this imparts the tragic sense that someone
has, out of desperation, failed to recognize his
or her own worth. This harmonizes well with
the mood of Eternal Sunshine, as the film offers
up exactly this sort of tragic situation in which
individuals are blind to their own worth: the
three people who we see using the memory-
removal procedure are all characterized as self-
destructive to varying degrees, with Clem’s
alcoholism, Joel’s depression, and Mary’s
insecurity and weakness of will making it all
too plausible that they would also engage in the
sort of harm to self that memory removal
involves. The film suggests that what they have
done is both sad and wrong; Kant’s moral theory
helps make this suggestion comprehensible.

126 Thinking Through Cinema: Film as Philosophy

VI. HARMING OTHERS THROUGH DECEIVING ONESELF

Watching the film, we do not simply feel bad
for Joel and Clementine because we suspect
they have harmed themselves in removing their
memories; we also naturally think that this pro-
cedure involves harming those who are erased
as well.24 Consider in particular the feelings of
sympathy that arise for Joel based on Clem’s
actions. Aside from worrying that he will harm
himself in choosing memory removal, viewers
of the film cannot help but think that Joel has
already been harmed by Clementine through
her trip to Lacuna. He certainly takes her
decision to remove memories of him as some-
thing of an insult, and we are inclined to agree.

There is a rather straightforward way of
understanding the nature of this harm, for we
see Joel’s confusion, sadness, and anger on the
screen as he learns about what Clementine has
done. He is made miserable by the news, and
the thought of removing this newfound misery
seems to be at least part of the basis for his
decision to undergo the procedure himself. We
saw earlier, though, that there are other classes
of harm that are trickier to make sense of: harms
that befall a person even though that person
does not experience the harms. I suggested that
Joel, Clementine, and Mary can be seen as
harming themselves in this way by undergoing
the memory-removal procedure; they harm
themselves through deprivation of the truth
regarding their previous relationships. I think
we can (and should) go one step further, how-
ever, and say that Clementine has not just
harmed herself but also harmed Joel in a way he
cannot experience.25 Just as in the case of unex-
perienced harm to self, this claim is initially
puzzling. It is clear enough that Clem has
harmed Joel in a very palpable way once he dis-
covers that she has had him erased, but it is a
significantly harder question whether he can be
said to be harmed even if he does not discover
what she has done.

We can better contemplate this possibility by
considering a scenario slightly different from
the one we saw in the film: imagine that
Clementine erased Joel, but Joel never came to
discover the erasure. (Perhaps he left to live in
another country before she underwent the pro-
cedure and he lost all contact with mutual
friends, family, and so forth.) Would it be right

to say that Clementine harmed Joel in her
actions? Opinions are likely to be divided here,
as we saw earlier that there are those (such as
many utilitarians) who find the idea of an
unexperienced harm nonsensical. Yet there are
also folks like Nagel, who plausibly suggest that
dismissing unexperienced harms may involve a
larger sacrifice to our ordinary intuitions and
commonsense than is initially obvious. If we
can legitimately say that betraying someone
behind their back involves harming them even
if they never discover the harm, it would
seem we should similarly be able to say that
Clementine’s actions harm Joel even if he never
finds out.

Granting that some harms are not necessarily
experienced, what is the nature of the non-
experiential harm perpetrated by Clementine?26

She has not exactly betrayed Joel, has she?
After all, one might think that choosing to
remove the memories of someone else is not
significantly different from throwing out their
old letters or deleting all their emails.27 Is it not
her right to remove mementos or even memo-
ries if she chooses? Perhaps, but here we may
be riding roughshod over morally relevant dif-
ferences between the case of an ex-lover burn-
ing letters and Clementine wiping all trace of
Joel from her mind. There is certainly a differ-
ence in degree between the two cases, and that
might be enough to make a moral difference,
but there is also something more: entirely wip-
ing out the memory of someone seems to mani-
fest a failure of respect that is distinct in kind
from merely discarding keepsakes.

On reflection, this sort of case appears to be
less like the tossing of old letters and more like
a genuine betrayal. Just as we might think that
someone who has misrepresented the memory
of someone else through slander has done him
or her a disservice, we can similarly say that
one who has removed all memory of someone
has also done a disservice to the person who has
been erased.28 Though the idea may initially
sound bizarre, it follows that we may have a
moral obligation to remember those we have
had close relationships with. Note that I did not
say we have a moral obligation to have fond
memories, or to like the person, for that would
clearly be a ludicrous demand. Rather, I am sug-
gesting that we are morally obliged to not distort
history through distorting our own historical record.

Grau Eternal Sunshine of the Spotless Mind and the Morality of Memory 127

Consider a drug that would revise one’s
memories such that all the memories of one’s
ex-spouse become both false and unflattering.
Many would rightly regard the taker of such a
drug as having done something that is not only
imprudent but also immoral. Although remov-
ing memories is not the same as distorting them,
the removal of all the memories of a person does
amount to a form of distortion: your mind
comes to have a falsified and thus distorted
perspective on one aspect of the world. Through
a voluntary “lie by omission,” the narrative of
your life has been, in part, fictionalized.

I said earlier that memory removal is disturb-
ing because it amounts to putting on “metaphys-
ical blinkers” that partially sever the connection
between one’s mind and the world—the mind
no longer reflects the world as accurately as it
did. There is symmetry in our values here: just
as we want our mind to accurately represent the
world, we also want the world to accurately
represent us.29 If I delete all my memories
of a person, I ensure that a part of the world no
longer represents that person at all, and it is
hard not to think that I have thus engaged in a
morally problematic form of misrepresentation.
If that person were to find out what I have done,
he or she would have the right to be offended.
Even if they do not find out, it is plausible to
think that they have nonetheless been harmed
by my actions. Though the degree of wrong-
doing may vary in accordance with my motives
(as in the case of slander or other forms of mis-
representation), even memory removal done
with the best of reasons can amount to a misfor-
tune for the person erased because it involves
this willful failure to represent the person
accurately.

I suspect some skepticism remains in many
readers for, despite the considerations above, a
duty to remember can seem like a very odd
thing for morality to demand.30 If we can free
ourselves from an overly narrow conception of
morality as nothing more than a collection of
abstract rules that regulate behavior towards
others, I think we can see that what I am sug-
gesting is not really that strange. The philoso-
pher and novelist Iris Murdoch can provide aid
here: she eloquently argued that at the core of
morality is a responsibility to do our best to get
things right, and this means not just to act
rightly but to perceive the world and other

people accurately—to “really look” and see
things as they actually are.

The authority of morals is the authority of truth, that
is of reality. We can see the length, the extension, of
these concepts as patient attention transforms accu-
racy without interval into just discernment. … Should
an unhappy marriage be continued for the sake of the
children? Should I leave my family in order to do
political work? Should I neglect them in order to
practice my art? The love which brings the right
answer is an exercise of justice and realism and really
looking. The difficulty is to keep the attention fixed
upon the real situation and to prevent it from return-
ing surreptitiously to the self with consolations of
self-pity, resentment, fantasy, and despair. … It is a
task to come to see the world as it is.31

If this characterization of “the authority of
morals” is correct, then I think it is quite reason-
able to conclude that we ought to do our best
not just to look, but also to not forget what we
have seen. Choosing to obliterate all trace of
someone else is the very opposite of the sort of
focused attention that Murdoch describes as
necessary for both love and justice.32 Maintain-
ing the ability to look back is just one part of our
larger responsibility to look at the world with
the clarity that morality requires.33 Eternal
Sunshine presents us with several characters
who have, for various reasons, chosen to evade
that responsibility, and the film effectively
cautions us against such escapism.

VII. MARY’S THEFT

Eternal Sunshine, unlike some science-fiction
films, is not in love with the new technology it
showcases.34 Quite the reverse: the memory-
removal procedure is presented as a tempting
but misguided and dangerous tool. I have
attempted to make sense of and defend the
pessimistic tone of the film toward this sort of
procedure. I have argued that undergoing
memory removal can amount to harming both
yourself and the person you have erased. You
harm yourself through depriving yourself of the
truth about your life and the world. You harm
the other person through a kind of misrepresen-
tation that is inevitable when you remove all
representation of that person. My hope is that a

128 Thinking Through Cinema: Film as Philosophy

consideration of the philosophical issues
involved has helped us to better understand and
justify the film’s technological pessimism and
its sense of tragic loss.

Our consideration of the philosophical issues
raised by Eternal Sunshine has put us in a better
position to defend the actions of Mary at the end
of the film. I mentioned earlier that she ulti-
mately decides to steal the medical records of
Mierzwiak’s patients and return them. From a
strictly utilitarian perspective it might seem
obvious that Mary is doing something morally
wrong in returning those files, mementos, and
audiotapes. Surely she will be causing many of
the previous patients pain and perhaps even
intense, prolonged suffering. Why, then, does she
think she is “in the right,” and why does the audi-
ence tend to sympathize with her actions? I have
been suggesting in this essay that there are a vari-
ety of ways in which we can think of the mem-
ory-removal procedure as causing significant
harm through deprivation. Mary is attempting to
undo that harm, and even if her attempt brings
with it some “harms that wound,” we (and she)
are inclined to think the suffering might well be
worth it. Although we do not get to see the full
results of her actions, the film suggests that her
goal is a worthy one, and the philosophers we
have considered have helped us acquire a fuller
understanding of why her actions may be justi-
fied despite the pain they will bring.35

VIII. CONCLUSION

I have tried to explain how the philosophical
resources provided by Nozick, Nagel, Kant,

Murdoch, and others put us in a position to
better understand why the scenario of Eternal
Sunshine is disturbing and morally problematic
in a way that cannot be fully accommodated by
traditional utilitarian thinking alone. Nozick’s
and Nagel’s insights suggest that we can legiti-
mately claim that memory removal involves a
conflict of values, one that results in harm to the
individual that goes beyond the sorts of harms
measurable in terms of utility. In addition, Kant
has given us reason to worry that voluntary
memory removal displays a lack of self-respect
that is harmful and perhaps immoral. Finally,
Murdoch’s moving vision of morality as requir-
ing accurate perception allows us to make sense
of the idea that memory removal can involve
not just harm to self, but also harm to others.

These considerations help us better under-
stand our response to the film: watching Eternal
Sunshine it is quite natural to feel uneasy
regarding the decisions Joel, Clementine, and
Mary make to utilize memory removal. It may
initially seem puzzling that we are led to feel
conflicted regarding the voluntary decisions char-
acters have made in order to pursue happiness.
(No one forces them, and they are choosing
memory removal in order to feel better, so why
should we feel ambivalence?) However, under-
standing the nature of their sacrifice, and the
manner in which it involves a kind of harm to
others as well as exploitation of the self, allows
us to make better sense of our emotional
response to the film. We can now see why the
depth of sadness evoked by the film is not
exhausted by a consideration of the bad conse-
quences and suffering we witness—the misfor-
tune the characters bring on themselves and
others is not always in the form of misery, but it
is misfortune just the same.

Filmmakers and novelists are often more
successful than philosophers at exploring the
nuances and complexity of our beliefs, desires,
and values. However, philosophy has a role in
helping us in the quest to make sense of and
interpret this complexity. Eternal Sunshine is,
among other things, a valuable philosophical
resource because it vividly illustrates the
poverty of the classical utilitarian perspective
through making us aware that moral reality is
significantly more complex than such utilitarian
theory can allow. In particular, the film shows
us that the harm caused by voluntary memory

FIGURE 2. Mary (Kirsten Dunst) leaving Lacuna, Inc. with
patients’ records in order to return them to those who have
had memories erased.

Grau Eternal Sunshine of the Spotless Mind and the Morality of Memory 129

removal cannot be satisfactorily understood
solely in terms of harms that are consciously
experienced. Philosophical argumentation is
required to make explicit these implicit lessons
of the film, and it has been my goal here to
utilize philosophical resources to do just that so
as to better understand both this remarkable
film and the philosophical issues it so
eloquently raises.

IX. POSTSCRIPT: THE PRESIDENT’S COUNCIL REPORT
ON BIOTECHNOLOGY

I have been discussing memory removal in the
context of the film Eternal Sunshine of the Spot-
less Mind. There, a procedure is utilized to wipe
out all memories of a previous relationship. I
suggested that the film gives us reason to think
that such a procedure is troubling, and I tried to
unpack this thought through relying on the
insights of various philosophers. My goal in the
above essay was primarily to argue that mem-
ory removal can be morally problematic in a
way that goes beyond its potential for bringing
about bad experiences. It does not follow from
this that memory removal may not, in some
cases, be justified. What does follow is that
even if it is justified, that does not necessarily
make it an unequivocally good thing—it may
instead be merely the lesser of two evils.

My qualification here is not simply for the
sake of academic accuracy, for the issues raised
by Eternal Sunshine are not as farfetched or
futuristic as some might think. As The New
York Times has reported, memory-diminishing
drugs have been given to those with posttrau-
matic stress disorder in an attempt to lessen the
horrible symptoms that can follow the witness-
ing of traumatic events.36 The drugs in question
do not exactly erase problematic memories, but
they do diminish them through blunting the
emotions connected to specific memories. The
President’s Council on Bioethics finds the
research disturbing enough to devote a chapter
on it in a recent report, and it cautions against
the use of such drugs.37 I want to briefly con-
sider their arguments and compare them with
the concerns I have raised regarding the mem-
ory removal exhibited in Eternal Sunshine.

The report raises many sensible worries
about the safety and effectiveness of such drugs,

but the heart of its philosophical argument
against memory removal seems to be threefold.
First, it claims that the happiness we seek from
memory removal would be a shallow simu-
lacrum of genuine happiness.

Yet it is far from clear that feelings of contentment sev-
ered from action in the world or relationships with other
people could make us truly happy. Would a happiness
that did not flow from what we do and say, usually in
association with others, be more than a simulacrum of
that happiness for which our souls fit us? (p. 208)

Second, the Council suggests that pursuing such
technology shows a failure to properly recog-
nize our limitations.

[B]y disconnecting our mood and memory from what
we do and experience, the new drugs could jeopard-
ize the fitness and truthfulness of how we live and
what we feel, as well as our ability to confront
responsibly and with dignity the imperfections and
limits of our lives and those of others. Instead of
recognizing distress, anxiety, and sorrow as appropri-
ate reflections of the fragility of the human life and
inseparable from the setbacks and heartbreaks that
accompany the pursuit of happiness and the love of
fellow mortals, we are invited to treat them as diseases
to be cured, perhaps one day eradicated. (p. 213)

Does not the experience of hard truths—of the
unchosen, the inexplicable, the tragic, remind us
that we can never be fully at home in the world,
especially if we are to take serious the reality of
human evil? (p. 229)

Finally, the Council worries that memory
removal involves a harmful tampering of one’s
personal identity.

But if enfeebled memory can cripple identity, selec-
tively altered memory can distort it. Changing the
content of our memories or altering their emotional
tonalities … could subtly reshape who we are, at least
to ourselves. With altered memories we might feel
better about ourselves, but it is not clear that the bet-
ter-feeling “we” remains the same as before. (p. 212)

[A]n unchecked power to erase memories, brighten
moods, and alter our emotional dispositions could
imperil our capacity to form a strong and coherent
personal identity. (p. 212)

130 Thinking Through Cinema: Film as Philosophy

[W]e might be often be tempted to sacrifice the accu-
racy of our memories for the sake of easing our pain
or expanding our control over our psychic lives. But
doing so means, ultimately, severing ourselves from
reality and leaving our own identity behind. (p. 234)

These and other considerations lead the Council
to issue a strong warning against such biotech-
nology.

Memory and mood-altering drugs pose a fundamen-
tal danger to our pursuit of happiness. In the process
of satisfying our genuine desires for peace of mind, a
cheerful outlook, unclouded self-esteem, and intense
pleasure, they may impair our capacity to satisfy the
desires that by nature make us happiest. (p. 269)

Regarding the Council’s claim that memory
removal might alter the patient’s personal
identity, the relevant issues seem to be not whether
one’s identity might be altered, but how it is
altered, and whether the change is for the best.
There is something morally problematic about
the idea of manipulating one’s mind, as our
consideration of Kant’s position showed, but
the problem is surely not just that one’s identity
has changed. After all, there are all sorts of
behaviors we can engage in that will, in some
sense, alter our identity.38 That a procedure
alters the self cannot, by itself, be a reason for
rejecting it.

Regarding the claim that pursuing such tech-
nology involves a denial of our limitations as
humans, the Council seems to be putting for-
ward a contentious theistic account of human
nature as inherently limited. (The Chair of the
Council, Leon Kass, is notorious for his conser-
vative, theistically-based positions, including
his initial rejection of in-vitro fertilization.)
Although I am rather skeptical of folks like the
“transhumanists” who giddily embrace the view
that technology will soon allow for a seemingly
unlimited increase in our abilities,39 I also find
it disturbing to encounter a government council
speaking of the natural “limits” of humanity and
advising against even the attempt to “feel at
home in the world.” A rejection of a particular
technology should be based on the actual dan-
gers it poses, not on the mere fact that it is new
and appears capable of reducing our limitations
in ways that would once have been thought of
as “unnatural.”

Regarding the claim that such drugs might
tempt us to accept a shallow or fake happiness
instead of the real article, I am sympathetic to
the Council’s point but wary of their rather
narrow conception of “genuine happiness.”
They seem in the end to suggest that genuine
happiness involves not feeling good but instead
simply being a good citizen.

Perhaps a remedy for our psychic troubles lies in the
rediscovery of obligations and purposes outside the
self—a turn outward rather than inward, a turn from
the healthy mind to the good society. And perhaps
the most promising route to real happiness is to live a
fully engaged life, as teachers and parents, soldiers
and statesmen, doctors and volunteers. (p. 267)

No doubt engagement with the community and
the society at large is often conducive to a
substantial and lasting sense of happiness;
however, it seems quite wrong to suggest that
true happiness is available only through such
engagement, or that such engagement will
necessarily bring contentment. Indeed, this
suggestion is pernicious if it implies that unless
such engagement brings happiness it is not
worth pursuing.

Rather than dismiss the idea that happiness
may be available through memory removal or
mood-altering drugs, I think it may be more
fruitful to instead point out that happiness is not
all that we care about in life. As Nozick’s
thought experiment shows, living a happy life is
not the only thing that matters—living a mean-
ingful life also has priority for most people. By
“meaningful life” I mean one in which a person
is able to realize his or her deepest values.40

Since many of us value more than simply happi-
ness, the meaningful life is not simply the
happy one. Memory removal (or memory dead-
ening through drugs) may inhibit our capacity
for meaningful lives, and caution is in order
when it comes to pursuing such technology. In
certain situations, however, medications or
technologies that blunt memories may instead
allow for a happiness and meaningfulness that
is not otherwise available. As Robin Henig
points out in The New York Times:

Without witnessing the torment of unremitting post-
traumatic stress disorder, it is easy to exaggerate the
benefits of holding on to bitter memories. But

Grau Eternal Sunshine of the Spotless Mind and the Morality of Memory 131

a person crippled by memories is a diminished
person; there is nothing ennobling about it. If we as a
society decide it’s better to keep people locked in
their anguish because of some idealized view of what
it means to be human, we might be revealing our-
selves to be a society with a twisted notion of what
being human really means.41

As I argued earlier, memory removal involves a
sacrifice because of the conflict between the
value we place on veracity and the value we
place on contentment. Such a sacrifice involves
a significant loss, but in certain circumstances
this loss may be outweighed by the gain made
in contentment, freedom, and psychic health.
Our duty to remember can be trumped by the
horribly debilitating effects of severe trauma
and, in such cases, it would be quite cruel to
deny relief to the person who is suffering.42

CHRISTOPHER GRAU
Department of Philosophy
Florida International University
Biscayne Bay Campus
North Miami, Florida 33181
USA

INTERNET: grauc@fiu.edu

1. I benefited from the discussions of an audience at the
University of North Florida, as well as the participants of
the 2005 Pacific Division Meeting of the American Philo-
sophical Association session hosted by the Society for the
Philosophical Study of Contemporary Visual Art. In add-
ition, I owe special thanks to Susan Wolf, Tim Mawson,
Murray Smith, Tom Wartenberg, Dan Callcut, Sean Greenberg,
Carlene Bauer, Chris Caruso, Susan Watson, Sean Allen-
Hermanson, Paul Draper, and Josh Oreck for comments on
earlier drafts of this essay.

2. It might seem obvious that any philosophical themes
in the film should be credited to the screenwriter, but in this
case it is not clear who gets the credit (or blame). The
screenplay was written by Charlie Kaufman, but based on a
scenario by Michel Gondry that is, in turn, based on an idea
by the French conceptual artist Pierre Bismuth. Also,
Kaufman’s original script is significantly different, with a
bleaker and more cynical ending. Presumably, either
Gondry or others pushed for the film to have a more
nuanced, romantic, and (cautiously) upbeat conclusion.

3. I will not be exploring the particular aesthetic and
film-theoretic issues raised by Eternal Sunshine, but that
certainly is not because I do not think they are worth explor-
ing—there is much that could be said about this notable
film. For example, others have pointed out that Eternal Sun-
shine seems to fit rather nicely within the genre of film that

Stanley Cavell has made famous with the label “Comedies
of Remarriage.” See Cavell’s Pursuits of Happiness: The
Hollywood Comedy of Remarriage (Harvard University
Press, 1981). Such films involve a separated couple
ultimately getting back together through rediscovering why
they fell in love in the first place. Eternal Sunshine follows
that pattern, but with the novel twist of memory removal
facilitating the “reunion.”

4. There is an interesting exception here: the technology
as it functions in the film (flaws and all) actually allows a
couple to reunite in a way that may not have been possible
otherwise. It is not clear whether this reunion is a good thing
(though many viewers, myself included, take it to be). Even
if a glitchy and incomplete memory removal brings about a
happy result in this particular case, however, this does not
warrant an acceptance of the technology in general.

5. In “Philosophy Screened: Experiencing The Matrix,”
Midwest Studies in Philosophy 26 (2003): 139–152, Thomas
Wartenberg has usefully pointed out that the idea that film,
a visual medium, can illustrate a philosophical claim is
itself worth questioning. Although I think this is indeed a
general issue worth pursuing, it seems to me the best way to
do this is to consider whether a particular film can in fact
succeed in such illustration. This is part of what I am under-
taking here. (When discussing The Matrix, Wartenberg ulti-
mately acknowledges that film can occasionally embody
philosophical argumentation as a form of “thought experi-
ment.” Though I do not dwell on this issue in this essay, I
think that a similar claim could be made regarding Eternal
Sunshine. Wartenberg’s comment that “the film actually
provides its viewers with a visual experience that is analo-
gous to [the protagonist’s]” seems equally applicable here
(p. 149). As with The Matrix, Eternal Sunshine begins by
forcing the viewer to enter an epistemic position similar to
that of the main character, and in the process “screens” a
thought experiment that can provide philosophical insight.)

6. One of the more insightful aspects of the film involves
its presentation of how the memory-removal technology is
actually implemented. Unlike most “sci-fi” films, which
offer naïve pictures of technological innovation being pur-
sued and employed by only the best and brightest (for
example, brainiacs in lab coats), we here see a much more
realistic portrayal of how this technology (if widely
marketed) is likely to be used: ordinary twenty-something
slackers perform the procedure with the same degree of
respect and competence that they would bring to developing
film at a one-hour photo lab. (Having worked at a one-hour
photo lab as an ordinary slacker I can speak with some
authority here. The manner in which Patrick unethically
keeps Clementine’s mementos for his own purposes is simi-
lar to the way in which some of my fellow employees would
make copies of photos they liked for their own use.) It is an
interesting question why most futuristic films fail to contain
this sort of realism regarding the manner in which tech-
nology is likely to be employed, though I will not further
pursue that issue here.

7. It should be noted that one of the many difficulties
with utilitarianism is that in situations such as the one we
are considering, utilitarian theory would seem to require
even the involuntary use of such technology if it would be
likely to maximize utility.

8. Robert Nozick, Anarchy, State and Utopia (New
York: Basic Books, 1974), p. 43.

132 Thinking Through Cinema: Film as Philosophy

9. Nozick, Anarchy, State and Utopia, p. 44.
10. Christopher Grau, “Bad Dreams, Evil Demons, and

the Experience Machine: Philosophy and The Matrix,” in
Philosophers Explore the Matrix, ed. Christopher Grau
(New York: Oxford University Press, 2005). My summary
of Nozick’s thought experiment here draws on my formul-
ation in that essay.

11. Charlie Kaufman, Eternal Sunshine of the Spotless
Mind (The Shooting Script) (New York: Newmarket Press,
2004), p. 111.

12. Ludwig Wittgenstein, Philosophical Investigations
(London: Macmillan Publishers, 1958), § 217.

13. Colin McGinn, Eternal Questions, Timeless
Approaches (New York: Barnes & Noble Audio, 2004).
McGinn further explores this theme in “Being and Know-
ingness” (unpublished manuscript, 2005).

14. There is another less direct way in which this proce-
dure brings on the threat of metaphysical isolation: if such a
procedure were actually possible, the ordinarily farfetched
skeptical worries that we may be radically wrong about our
past become much less farfetched and much more worri-
some. No one could be sure that they had not in fact had
large portions of their lives erased at some earlier point.
(This might be grounds for doubting that a utilitarian
defense of such technology could ever be feasible, for it is
hard to see how this sort of worry could be eliminated if the
existence of the procedure became widely known.) The
worry here is related to the skeptical worries regarding
artificial memories that are raised by such films as Blade
Runner (Ridley Scott, 1982) and Total Recall (Paul Verhoe-
ven, 1990).

15. Thomas Nagel, “Death,” in The Metaphysics of Death,
ed. John Fischer (Stanford University Press, 1993), p. 64.

16. Steven Luper, “Death,” in the Stanford Encyclopedia
of Philosophy, available at .

17. Nagel’s primary concern is to make sense of the idea
that death harms the one who dies. Acknowledging death as
a deprivation brings with it additional metaphysical diffi-
culties (e.g., when is one deprived?) that do not confront the
case of memory removal.

18. There is another way in which we can see this tech-
nology as involving harm to oneself: on some philosophical
accounts of the self (such as John Locke’s), personal iden-
tity consists in the continuity of and connections between
memories. Put bluntly, on this sort of account you just are
your memories. Such approaches to identity bring with
them the consequence that a loss of memories is quite
literally a loss of the self. Memory removal becomes “self-
destructive” as a matter of definition! (I do not find such
a criterion for personal identity particularly persuasive, so
I do not dwell on this issue here.)

19. Immanuel Kant, “Grounding for the Metaphysics of
Morals,” in Ethical Philosophy, trans. James W. Ellington
(Indianapolis: Hackett Publishing, 1983), pp. 36–37, § 429.

20. Immanuel Kant, Lectures on Ethics, ed. Peter Heath
and J. B. Schneewind, trans. Peter Heath (Cambridge
University Press, 1997), p. 125.

21. James Rachels, The Elements of Moral Philosophy
(Boston: McGraw-Hill, 2003), p. 136.

22. It is a further question whether Kant is right to clas-
sify all forms of rehabilitation as manipulation. It seems that
Kant should have distinguished between those methods that

involve trying to reason with the agent from those that use
nonrational means and coercion to induce change.

23. Kant, Lectures on Ethics, p. 127.
24. Mary’s case is significantly different, of course, as

she has Mierzwiak’s consent and encouragement.
25. The thought that she could harm Joel through erasing

him may have been her primary motivation for undergoing
the procedure. She proudly admits at one point that “I’m a
vindictive little bitch truth be told.” However, their mutual
friend Carrie tells Joel: “What can I say Joel … She’s
impulsive … She decided to erase you almost as a lark,” so
the degree of intended malice in the act is not entirely clear.

26. I do not here explore a specifically Kantian account
of the harm done to others, but perhaps such an account
could be defended. It is not clear that a maxim involving the
desire to erase memories of someone could be universal-
ized, and it also seems plausible to suggest that choosing to
erase the memories of someone embodies a failure of
respect for the person erased. Ken Rogerson has suggested
(in conversation) that utilizing memory removal may vio-
late Kant’s absolute prohibition on lying, for in erasing all
trace of someone you intentionally place yourself in a posi-
tion in which, if asked, you are bound to say false things
about your own history and the person you have erased.
There is also the obvious possibility that memory removal
will cause one to disregard a Kantian prohibition on break-
ing promises: if you cannot remember your promises, you
cannot possibly be sure you will keep them.

27. In a recent New York Times article, Anna Bahney
points out that thanks to the dominance of email and digital
photography, these days the mementos of a love gone sour
can be “expunged with brutal efficiency” (“Zapping Old
Flames Into Digital Ash,” April 4, 2004).

28. I am here focusing on the harm done to the person
you have erased, but given the way the memory removal
functions in the film, other questions come up regarding
your duties to those who are constrained by your actions. Is
it moral of you to request that all mutual friends refrain
from mentioning the procedure you have undergone? Ask-
ing them to act as though your relationship never occurred
could, in certain circumstances, amount to imposing a very
significant burden.

29. If we feel that part of our nature is reprehensible,
however, we may embrace the opportunity to hide this
fact. (Thus Mierzwiak feels no need for Mary’s mind to
accurately reflect his full nature.)

30. In his book The Ethics of Memory (Harvard Univer-
sity Press, 2002), Avishai Margalit argues that we have an
ethical rather than a moral obligation to remember others.
He aligns ethics with what Bernard Williams has called
“thick relations” with those we care about, while morality is
relegated to “thin relations” with those we are less con-
nected to. He thinks the realm of ethics is optional in a way
morality is not, for he concludes that our obligation to
remember is conditional on our desire to be involved in
caring relations with others. Although I think there is
something insightful in the distinction between ethics and
morality (as well as between thick and thin), I am not as
confident as he is that the line between the two can be
clearly drawn, or that one is optional in a way that the other
is not. Accordingly, I do not hesitate to speak of special
moral obligations to those we care about. (My arguments
for why we have an obligation are rather distinct from

Grau Eternal Sunshine of the Spotless Mind and the Morality of Memory 133

Margalit’s, though I do think much of what he says about
the relationship between care and memory is compatible
with my account. In particular, I suspect Iris Murdoch’s
emphasis on focused attention could be reformulated in
terms of “care” as he uses that term.)

31. Iris Murdoch, The Sovereignty of the Good (London:
Routledge & Kegan Paul, 1985), pp. 90–91.

32. It is rather ironic that Kate Winslet has played both
Iris Murdoch (in the film Iris (Richard Eyre, 2001)) and
Clementine, a character who exhibits the reluctance to
“really look” that Murdoch criticized. (It is also sadly ironic
that Iris Murdoch herself came to suffer from Alzheimer’s,
and thus gradually lost just the sort of perceptual acuteness
she felt was so important for love and justice.)

33. My utilization here of Murdoch the moral particularist
alongside Kant the universalist may strike some readers as
bizarre. It certainly would be bizarre (and perhaps incoherent)
if I were urging that we accept Kant’s entire moral theory.
However, I think we can benefit from Kant’s insights regard-
ing the problematic nature of self-mutilation without commit-
ting ourselves to his overall conception of morality, just as we
can benefit from Murdoch’s insights regarding focused atten-
tion and the connection between love and justice without
accepting all that she says on the topic of morals.

34. In an insightful article, Andrew Light has argued that
the film The Conversation (Francis Ford Coppola, 1974)
can be seen as supporting what Light calls a “substantive”
thesis regarding technology. See “Enemies of the State,” in
Reel Arguments: Film, Philosophy, and Social Criticism
(Boulder, CO: Westview Press, 2003). This thesis (originat-
ing in the works of Martin Heidegger, Michel Foucault, and
others) involves the view that technology is not “value free”
or morally neutral, but has an inherently ethical dimension.
In particular, Light suggests that the surveillance techno-
logy employed in The Conversation is presented as corrupt-
ing its users. Light explains this corruption in terms of the
alienation that is inevitably imposed by the technology: by
its very nature the surveillance technology tends to objectify
the person being surveyed. I think something similar could
be said in favor of a “substantive” interpretation of Eternal
Sunshine. The film presents the actual employment of the
memory-removal technology as inevitably involving the
depersonalization and manipulation of the patient. This in
turn seems to lead to other moral infractions great and
small: Patrick steals both underwear and Joel’s girlfriend,
Stan utterly fails to show respect for Joel as a patient,
Mierzwiak uses the technology to evade responsibility for
his affair with Mary, and so forth. Just as The Conversation
suggests that surveillance technology corrupts the character
of those utilizing it, Eternal Sunshine seems to suggest that
memory removal technology is also far from morally
neutral and brings with it a problematic attitude of objec-

tification that infects those charged with controlling the
technology.

35. I say “may be justified” here because it is possible
that the return of the files will indeed cause so much suffer-
ing that it will counterbalance the good accomplished
through undoing the harm of deprivation caused by memory
removal. There are also questions of patient consent that
may be morally relevant. I do not want to deny these possi-
bilities and complications—I just want to suggest that the
fact that we seriously consider the judgment that Mary’s
actions are justifiable shows us that both harms that wound
and harms that deprive need to be recognized here. Without
the recognition that there has been harm through depriva-
tion, Mary’s actions become obviously unjustifiable, yet
when we watch the film we do not find her to be so obvi-
ously in the wrong.

36. Robin Marantz Henig, “The Quest to Forget: Drugs to
Prevent Painful Memories,” The New York Times April 4, 2004.

37. This report, Beyond Therapy: Biotechnology and the
Pursuit of Happiness, is available at . Future references will be to
page numbers only.

38. Indeed, it is hard to think of activities that do not, in
some sense, change who we are. The incredible popularity of
“self-help” books is a testament to the degree to which we not
only permit, but seek out, opportunities to alter the self.

39. Resources regarding transhumanism can be found on
the web at . Nick Bostrom
is perhaps the best known transhumanist who is also an aca-
demic philosopher.

40. It would take us too far afield for me to go into
much more depth here, but it should be pointed out that
my characterization of meaningfulness is meant to be
neutral between “subjective” and “objective” accounts of
meaningfulness and value. Although I am sympathetic to
objective accounts, for my purposes here the subjective
account can suffice. For an insightful discussion of why
subjective accounts of meaningfulness are problematic,
see Susan Wolf’s essay “The True, the Good, and the
Lovable,” in Contours of Agency: Essays on Themes from
Harry Frankfurt, eds. Sarah Buss and Lee Overton (MIT
Press, 2002).

41. Henig, “The Quest to Forget.”
42. Note also that some of the objections I raised to the

memory-removal technology that appears in Eternal Sun-
shine may not apply to cases involving posttraumatic stress
disorder and memory-blunting drugs. There may be no sig-
nificant harm inflicted on others through the use of such
drugs. Also, to the extent that the drugs blunt emotional
tonalities rather than the factual content of memories, the
proper use of such drugs may not threaten the value we
place on knowing the truth.

Feature 08

43 Film Matters Winter 2012 ➜

Mainstream Mulvey:
Eternal Sunshine of the Spotless
Mind and (500) Days of Summer as
Alternatives to the Feminist
Avant-Garde
By Brenna Claire Williams
/
k e y w o r d s : Laura Mulvey, female gaze, gender in film,
scopophilia, romantic comedy, independent film

Gender politiCs in MainstreaM
Hollywood films have been at the center of
cinematic debates at least since 1975 when
laura Mulvey wrote “Visual pleasure and
narrative Cinema.” in the essay, Mulvey
criticizes film for its subjectification of women
and affirmation of dominant patriarchal
constructs and ideals. Mulvey posits that
the only place where these values can be
eradicated is in radical avant-garde cinema,
which “provides a space for cinema to be
born which is radical in both a political and
aesthetic sense” (Mulvey 716). However, since
Mulvey first wrote her essay, and especially
in the last decade, Hollywood films have
created a new kind of female character that
is able to hold the narrative and visual power
that Mulvey said wasn’t possible. As seen in
the films Eternal Sunshine of the Spotless Mind
(Gondry, 2004) and (500) Days of Summer
(Webb, 2009), the female love interest in
romantic comedies can not only be equal,
but can also hold more power than her male
counterpart. While these films were not
widely released blockbusters, they were each
distributed by the independent branches
of major studios and earned decent box-
office revenue, over $72 million and $60
million respectively (Box Office Mojo “‘eternal
Sunshine’”; “‘(500) Days’”), providing a
compromise between the mainstream cinema
that Mulvey loathed and the avant-garde
cinema that wider audiences overlook.

b e l o w Clem waves at Joel at the train station

Williams.indd 43 2/13/13 5:02:32 PM

44

Film Matters Winter 2012

Feature 08 Brenna Claire Williams

As seen in the films Eternal Sunshine of the Spotless Mind and (500) Days of
Summer, the female love interest in romantic comedies can not only be equal, but
can also hold more power than her male counterpart.

a b ov e Joel waits for Clem to come home a b ov e Clem and Joel on icy lake

a b ov e Joel alone on ice a b ov e Joel and Clem on tV

nontraditional narrative styles,1 self-awareness
and play with visual coding, and disregard
for and inversion of patriarchal gender
roles make these films realizations of what
Mulvey said film could be in a format that can
communicate positive depictions of women to
a wider audience than the avant-garde allows.

one of Mulvey’s most scathing criticisms
of mainstream cinema is the divide “between
active/male and passive/female” (Mulvey
719) with a woman’s “presence [tending] to
work against the development of a storyline”
(Mulvey 719). Eternal Sunshine and (500)
Days work against this trend by making
their female characters the ones whose
desires and actions further the plot while
men take reactionary roles, particularly
when considering the directions of their

relationships. Clementine (Clem) in Eternal
Sunshine takes control in almost every aspect
of her and Joel’s relationship and, similarly,
Summer is the one who begins, defines, and
ends her relationship with tom.

throughout Eternal Sunshine, Clem
dominates her relationship with her strong
will and opinions. Both times she and Joel
meet – first on the train from Montauk
and later on the beach – Clem speaks first,
pursuing Joel while he only looks at her from
afar. Clem’s desires drive the beginning of
their relationship both times: initially when
she invites him to her apartment and later
when they meet again and she leads Joel on
their excursion into the beach house. Joel
blindly follows her lead, listening to her when
she tells him to call her and risking his life

on an icy lake where she is able to push him
around because he is also physically passive.
additionally, Clem ends her relationship
by walking out on Joel and ignoring his
attempts to get her back. Clem is anything
but the “passive/female” character Mulvey
attributes to Hollywood films.

similarly, summer is in complete
control of her relationship with tom from
beginning to end. Like Clem in Eternal
Sunshine, summer follows her whims, even
moving across the country because she
wants excitement, while Tom, like Joel,
leads a relatively static, passionless, and
disappointing life. tom doesn’t admit that
he has feelings for summer, even after she
straightforwardly asks him. she is the driving
force behind their kiss in the copy room

The most interesting parallel between these two films is the fact that they are both
nonlinear, a characteristic that signals a denial of male power (Silvey 139). While
both films follow male protagonists, the nonlinear narratives make it difficult for
audience members to situate themselves within their lives and the world of the
film, which foregrounds the women’s importance in them.

Williams.indd 44 2/13/13 5:02:36 PM

Feature 08

45 Film Matters Winter 2012 ➜

Mainstream Mulvey

Perhaps the thing the films communicate most strongly is the idea that strong,
unique, and independent women are unforgettable. They are something to hold
onto, something that, despite what Dr Mierzwiak tells Joel, you will miss, for better
or for worse.

a b ov e Clem and Joel at the flea market a b ov e Joel and Clem at the movies

b e l o w Joel and Clem on the couch b e l o w Clem sticks up for Joel

and defines their relationship when they
kiss again at iKea. summer’s view of their
relationship wins even after tom is hurt
at the bar and tries to assert himself. Like
Clem, summer ends her relationship with
tom, operating as an active character from
the beginning to the end.

this active/passive division, according
to Mulvey, goes beyond gender power plays
and is found even at the core of the narrative
structure of films, with men “forwarding
the narrative, making things happen” and
serving as “main controlling [figures]”
around whom films are structured (Mulvey
720). The most interesting parallel between
these two films is the fact that they are both
nonlinear, a characteristic that signals a
denial of male power (Silvey 139). While
both films follow male protagonists, the
nonlinear narratives make it difficult for
audience members to situate themselves

within their lives and the world of the film,
which foregrounds the women’s importance
in them. in Joel’s case, Clem permeates every
part of his mind and memory, and in tom’s
case, Summer defines 500 Days of his life.

Eternal Sunshine opens on the morning
following Joel’s procedure and situates him
and the audience in his bleak apartment
and life, both devoid of the color Clem
brings to them. Here, the film adds a visual
component that works with the nonlinear
narrative structure in order to signal to the
audience how important Clem is in the film.
the colors are neutral until Joel sees Clem
in Montauk, which signals her as a strong
visual and narrative element since she stands
out from across a desolate beach. By the end
of the nonlinear film, the audience knows
that it is Clem’s imagined last words to Joel
that brought him to the beach, affecting him
and controlling the plot from the beginning.

Clem’s haunting personality wouldn’t have
been as effectively communicated if this
scene had appeared at the end of the film, as
it would have had it played chronologically.

Throughout the rest of the film, Clem
not only does what she wants to do, but also
affects the narrative structure of the film
by operating as the catalyst for most of the
things that Joel does, both while conscious
and during his procedure. She is the first to
get Joel erased from her mind, doing so in
an impulsive pursuit of her desires. in his
mind, he tells her that he is only erasing her
because she did it to him first. Within Joel’s
mind during the procedure, it is his mental
manifestation of Clem who has the idea
of going “off the map” (Eternal Sunshine of
the Spotless Mind) when Joel doesn’t have a
plan. Clem’s hold on Joel’s life, while they’re
together, after they’ve split, and even while
he is unconscious, continually reverses

The films are able to create a safer screen space for their main female characters
by noting visual and narrative elements like fetishism, scopophilia, and the
gender constructs to which those in a patriarchal society are accustomed.

Williams.indd 45 2/13/13 5:02:41 PM

46 Film Matters Winter 2012

Feature 08 Brenna Claire Williams

b e l o w Clem overpowers Joel
on the beach

a b ov e Joel from afar at the beach

b e l o w the use of split-screen highlights the interplay
between reality and Tom’s expectations

Mulvey’s assertion of male power
throughout the film.

likewise, the nonlinear narrative style
of (500) Days serves to deconstruct male
power, but in a different way than it does in
Eternal Sunshine. From the opening credits,
the focus on and importance of both summer
and Tom is introduced by their parallel
montage sequences, visual signals of the
fact that it is a story about both of them,
not just Tom. The audience’s position in
the story is guided both by a narrator and
title cards with the day of summer each
sequence depicts. While Eternal Sunshine
flows between present, reality, subconscious,
and past without pausing, (500) Days always
tells viewers where they will be in Tom and
summer’s relationship, with everything after
Day 290 depicting the aftermath with title
cards in shades of gray. the human mind
does not replay memories back accurately
or in order, something this film’s narrative
reflects, which gives the audience insight into
the way Tom sees the relationship. At first

tom and his audience miss the warning signs
of Summer’s unhappiness, but this gives
way to the realization and regret of missing
them when both Tom and the audience
replay memories more accurately during
Tom’s healing process. The film poignantly
and self-consciously highlights this interplay
between reality and Tom’s expectations by
splitting the screen between them when
tom goes to summer’s party. the days of
summer that the audience sees, like those
in Eternal Sunshine, highlight the power that
summer has not only over tom’s memory,
but also in reality. Like Clem in Joel’s life,
summer impacts tom’s life even when
they are not together. Before she and tom
start seeing each other, her attendance of
the company karaoke party spurs tom to
attend, and after they split, her insistence on
Tom’s ability as an architect pushes him to
pursue it again. summer, like Clem, serves
as a counterexample to Mulvey’s points by
making things happen and controlling much
of the narrative.

Mulvey spends much of her article
focused on the male gaze and accuses men
of “taking people as objects, subjecting them
to a controlling and curious gaze” (Mulvey
717) and deriving pleasure from looking,
which gives them power. Both Eternal Sunshine
and (500) Days consciously play with this
“scopophilia” by sharing the gaze with their
female characters and by robbing the men
of the pleasure of looking. Mulvey assigns
erotic pleasure to the male gaze, which the
films also take up and critique in different
ways: Eternal Sunshine makes male pleasure
something that Joel is ashamed of and (500)
Days has summer’s desires trump tom’s.
Unlike the films that Mulvey criticizes, these
films punish their male characters instead of
their female characters (Mulvey 722).

Clem actively looks both times she and
Joel meet, and the film draws attention to it.
When Joel and Clem are in the cafe at the
beginning of the film, Joel looks over at her.
However, she looks back, embarrassing him
and forcing him to turn his gaze elsewhere.

Williams.indd 46 2/13/13 5:02:46 PM

Feature 08

47 Film Matters Winter 2012 ➜

Mainstream Mulvey

above The film breaks Summer down into each part that
Tom loves about her. Here, Tom talks about her smile above Tom talks about how much he loves Summer’s laugh

above Summer is the driving force behind their kiss in the copy room above Summer initiates the first conversation with Tom

She is able to look at him even though he is
aware of it while he cannot look once robbed
of his voyeuristic pleasure, which suggests
that her gaze is more powerful than his or,
at the very least, that she is less ashamed of
it. at the train station, Clem is once again
aware that Joel has noticed her and waves at
him, effectively acknowledging his role as a
voyeur and her role as a gazing and aware
woman. on the train, Clem approaches Joel
and says, “i’ve seen you,” a sentiment that is
echoed when they meet for the second time
and she sits down next to him at the beach
and says almost the exact same thing. (500)
Days reflects this assertion of the female gaze
over the male gaze precisely when, at the end
of the film, Autumn tells Tom that she has
seen him before even though he hasn’t seen
her. Both times Clem and Joel meet, the film
highlights her agency, her gaze, and its power
while diminishing his.

summer’s gaze in (500) Days is much
more equal to Tom’s and goes a step beyond
Clem’s gaze by allowing the audience to
share in it. Mulvey says that women are,
“indispensible [elements] of spectacle in
normal narrative film” (Mulvey 719), but
(500) Days takes the opportunity to make
both Summer and tom elements of spectacle
and the gaze of the other when they sing
karaoke. When summer goes to sing, the
audience gets a shot of tom looking at her
with pleasure followed by a point of view
shot from his perspective. similarly, when
Tom sings, the film follows the same shot
progression and allows the audience to see

through summer’s eyes and share in her gaze
while tom makes a spectacle of himself.
Additionally, the film makes Tom the object
of a female audience’s gaze by having his
sister’s friends stare at him. While it isn’t a
fully realized erotic gaze, this example serves
as another acknowledgement of the gaze and
gives it to women. later, when tom gives
summer an architectural tour of los angeles,
the audience shares in another point-of-view
shot. However, it isn’t clear whose point
of view they are given, implying a shared
ownership of the gaze and situating the
couple, once again, in a gazing relationship
that is equal.

in the same vein as the scopophilic
gaze, Mulvey asserts that, “conventions of
mainstream film focus attention on human
form” (Mulvey 718). Both films directly
reference fetishistic scopophilia, which
Mulvey defines as the breaking down of the
female form to overvalued parts (Mulvey
720). In Eternal Sunshine, Joel muses about
Clem’s sweatshirt and wonders how he
could be “drawn to someone’s back.” Joel
also draws Clem in his sketchbook almost
as a hobby, pulling attention to the male’s
attention to the female form as well as the
way he constructs his own view of it. Joel
doesn’t photograph Clem, but draws her
from his mind’s construction. The majority
of what the audience sees of Clem is within
Joel’s mind, exposing them even further to
this construction.

on top of parodying the visual
construction of women in Hollywood films,

Eternal Sunshine presents Clem, through her
emotions, as a caricature of more typical
female characters, revealing the ridiculous in
those characterizations. Clem complains to
patrick that she feels and looks old, a concern
real and fictional women share. One of Joel’s
favorite memories of Clem is one in which
she asks if she is ugly and tells him about a
preoccupation she had as a child with being
pretty. While Clem is an extremely strong
female character, these instances show that
even the strongest woman cannot escape
the always damaging and traumatizing
effects of patriarchy’s demands, and the film
demonizes those demands by imposing them
on an otherwise relatable and admirable
female character. By vilifying patriarchy in
this way, instead of ignoring and escaping
it the way the avant-garde would try to
do, Eternal Sunshine draws the audience’s
attention to the evils that Mulvey sees in
patriarchy, which does more to educate it
about the problems of conventional film in a
palatable form.

(500) Days also deliberately parodies the
breaking down of women into parts and
desires. When introducing summer, the
narrator breaks Summer down by discussing
her height, weight, and shoe size and
whether each is average. later, tom talks
about the things he loves about Summer,
including her smile, hair, knees, lips, and
birthmark. The film breaks her down into
those parts by both visually depicting them
and through voice-over while noting almost
nothing about her personality. This echoes

Williams.indd 47 2/13/13 5:02:52 PM

48 Film Matters Winter 2012

Feature 08 Brenna Claire Williams

the idea presented and simultaneously judged
by Eternal Sunshine when patrick falls in love
with Clem’s unconscious body and not who
she is. By using the same visuals of summer’s
parts with a different voice-over of tom
saying that he hates them, the film highlights
the simultaneous adoration for and loathing
of the female form that Mulvey discusses
in her essay (Mulvey 721). In an interesting
twist, (500) Days also alludes to female
pleasure in looking by the breaking down
of the male form when tom’s sister talks
about “Brad Pitt’s face and Jesus’ abs,” which
consciously highlights female fetishism of the
male form – something that Mulvey ignores.

Eternal Sunshine and (500) Days also share
their acknowledgement of the idea of an
ideal female form, but in different ways,
with Eternal Sunshine deconstructing the ideal
and (500) Days presenting it as something
that every male holds in his mind. Eternal
Sunshine presents an image of Clem that
is anything but the idealized Hollywood
construct of the female form. she often
dresses in a baggy orange sweatshirt or in
multiple heavy layers, drawing away from
her figure and intentionally playing down
her beauty instead of accentuating it the way
most films would. The film draws attention to
the purposeful construction of Clem’s look,
changing her hair to unnatural hair colors
and assigning her “personality in a paste.”

Conversely, (500) Days presents its
audience with a specimen of the perfectly
beautiful and classic woman. The narrator
tells the audience that summer averages 18.4
double takes on her daily commute, a part
of the phenomenon the narrator refers to as
the summer effect. the summer effect,
the film claims, is something that every man
has encountered by falling under the spell
of a beautiful woman. Interestingly, the
film then attempts to deconstruct the idea
of the perfect woman, both when Tom and
summer discuss each other’s imperfections
and when each of the male characters are
shown, as if on 35 mm film, talking about
their ideal woman. Notably, Tom doesn’t
have a conclusion and paul admits that the
perfect woman simply isn’t a reality.

The films share a high degree of self-
awareness, communicating a sense of
purpose in everything they do, especially
in their depictions of women. this self-
awareness in Eternal Sunshine is probably
best exemplified in Joel’s memory when
he and Clem appear on the television in
their apartment. At that moment, the film
acknowledges an awareness seen throughout
the film that Clem is a cinematic or, in this
case, televisual construction of femininity
like any other woman on film or television.

(500) Days contains a similar moment that
functions almost as a parody of Mulvey’s
ideal avant-garde in addition to the
acknowledgement of film as construction.
tom goes to the movie theater after he and
Summer split up and projects his feelings
onto the screen in the form of an art film.
the scene displays the characters in (500)
Days as the movie characters they are while
almost taking a jab at Mulvey’s elevation
of the avant-garde as the only cinema that
can communicate truth. (500) Days itself
is certainly more enjoyable than the avant-
garde film within it.

The final way that Eternal Sunshine and
(500) Days reinterpret their depictions of
gender is by inverting or calling attention to
the dominant patriarchal gender roles that
Mulvey sees in film (Mulvey 716). At times,
each film switches the gender roles within its
couple, emasculating and feminizing the men
while empowering the women. additionally,
the female characters themselves call
attention to the way films portray normal
female characters by embodying two sides
of the spectrum of contemporary female
characters.

Eternal Sunshine invests Clem with multiple
stereotypically masculine personality
traits, such as aggression, which she is
able to play out upon a passive, quiet, and
femininely coded Joel. The first time they
meet in the film, Clem makes jokes about
seducing him, an idea that clearly makes
Joel uncomfortable. Clem talks about living
her life to the fullest, aggressively taking
advantage of the opportunities the way
male characters often do, while Joel lets life
happen to him. When Joel gets home, he
anxiously waits by the phone trying to wait
an appropriate amount of time to call Clem
the way a teenage girl might do in another
romantic comedy. this image is mirrored
in Joel’s last memory of Clem, when he sits
in the same chair waiting for Clem to come
home, nagging her “like an old lady” (Eternal
Sunshine) when she arrives. This interesting
juxtaposition between teenage excitement
and elderly patience ties Joel to a range of
feminine traits. While Clem is at times used
as a caricature of female emotions, Joel is
anything but stoic throughout the film, often
getting upset and crying in his childhood
memories, where Clem is notably collected
and in charge, all the way up to his present-
day adulthood.

similarly, (500) Days goes out of its way
to consciously make summer aggressive
and self-assured, coding her as masculine,
while making tom idealistic and unsure,
coding him as feminine. The beginning
narration makes the audience aware that

the film is “not a love story” ((500) Days) in
the traditional sense by painting Tom as a
hopeless romantic – The Graduate’s (nichols,
1967) fault, we’re told – and Summer as
unattached and guarded because of her
parents’ divorce. tom falls for summer
at first sight and overanalyzes everything
she says and does to him before they start
seeing each other, which feminizes him
the same way Joel’s anxiety does in Eternal
Sunshine. Summer believes that love isn’t real
and that it’s better to be guarded and safe,
traits usually assigned to male characters.
summer and tom want different things, and
in the end summer’s masculine desires for
detachment dominate tom’s stereotypical
feminine desire for attachment and mutual
dependence.

The film also constantly discusses the
adoption and construction of gender roles.
the narrator tells the audience that “there
are only two types of people in the world:
women and men,” an idea that (500) Days
plays with continually. in the diner, summer
compares herself and tom to sid Vicious
and nancy spungen, the rock star and
the girlfriend he stabbed to death. Tom
notes this and automatically assumes that
summer is implying he, as the man, is sid
and she, as the woman, is nancy. However,
summer says that she is sid, a conscious
reversal of gender roles and an outing of
the assumptions about them. In the first
karaoke bar scene, Summer discusses her
dating policy with tom and McKenzie,
saying that she doesn’t like attachment and
she is happy being independent. McKenzie
responds by asking if she’s a lesbian, which
communicates an awareness of the cultural
assumption that there is something Queer
about strong and independent women. When
summer says that she wants to save serious
relationships for later, McKenzie goes a
step further and drunkenly accuses her of
being a man. Summer, who the film again
makes masculine through her ideological
alignments with McKenzie, serves as a foil
to Tom, who the film feminizes through his
ideological opposition to summer and his
traditional male friend. interestingly, when
tom adopts summer’s nonchalant, non-
labeled view of relationships, McKenzie
tells him that he sounds gay, even though
he attributed the same traits to masculinity
when he saw them in Summer. By both
making the audience aware of cultural
conceptions of inherent gender traits and by
inverting those traits in its two protagonists,
(500) Days is able to communicate the
impermanence and hypocrisy of the gender
constructs of the patriarchal order that the
audience has undoubtedly accepted, for the

Williams.indd 48 2/13/13 5:02:52 PM

Feature 08

49

Mainstream Mulvey
Film Matters Winter 2012

Author Biography
Brenna Claire Williams is a 2012
Television Studies graduate of the
University of Notre Dame. Her research
focused on religion and the media,
culminating in her senior honors
thesis about South Park’s treatment of
minority religions. She now resides in
Washington, DC where she has written
for NPR and works for NBC News.

Mentor Biography
Jim Collins is a professor of film and
television and a concurrent professor
in English. He specializes in media
theory, contemporary narrative, and
digital culture. His most recent book is
Bring on the Books for Everybody: How
Literary Culture Became Popular Culture
(2010).

Department Overview
Majors in film, television, and theatre at
Notre Dame get more than an excellent
liberal arts education – they also have
the advantage of being in a department
with a high degree of student-faculty
interaction, a wide variety of hands-on
courses, and state-of-the-art work and
performance facilities.

➜
➜

most part, without question. This once again
communicates something Mulvey would
likely agree with in a more palatable way
than the radical avant-garde.

Overall, the films Eternal Sunshine of the
Spotless Mind and (500) Days of Summer are
able to formulate some of Laura Mulvey’s
main critiques of mainstream film for an
almost mainstream audience. The films
are able to create a safer screen space for
their main female characters by noting
visual and narrative elements like fetishism,
scopophilia, and the gender constructs to
which those in a patriarchal society are
accustomed. Clementine and summer are
active, assertive, aggressive, and powerful,
which affects the narrative structure of
their stories and the lives and actions of the
men who love them. Both films parody or
deny male pleasure and the male gaze by
sharing or completely handing power over
to the women in many cases and depicting
them in a positive light that even Mulvey
might approve. While the films are not
the perfect realization of feminist cinema,
they are palatable to a large audience, and
being able to get a feminist message to a
wider audience is a very large first step in a
positive direction. I would much rather be
Clementine or summer than any disney
princess or Marilyn Monroe character – and
let’s be honest, they get to wear some amazing
clothes – because they communicate that it
is okay to have opinions, be different, and to
be whimsy incarnate. Perhaps the thing the
films communicate most strongly is the idea

that strong, unique, and independent women
are unforgettable. They are something to
hold onto, something that, despite what dr
Mierzwiak tells Joel, you will miss, for better
or for worse.
/ e n d /

Works Cited

(500) Days of Summer. Dir. Marc Webb. Fox
Searchlight Pictures, 2009. DVD.
/
Box Office Mojo. “‘(500) Days of Summer’ (2009).” 3
Dec. 2011. Web.
/
Box Office Mojo. “‘eternal sunshine of the spotless
Mind’ (2004).” 3 Dec. 2011. Web.
/
Eternal Sunshine of the Spotless Mind. dir. Michel
Gondry. Focus Features, 2004. Film.
/
Mulvey, laura. “Visual pleasure and narrative
Cinema.” Critical Visions in Film Theory: Classic
and Contemporary Readings. ed. timothy Corrigan.
Boston: Bedford/St. Martins, 2011. 713–25. Print.
/
Silvey, Vivien. “Not Your Average Love Story: Film
techniques in ‘Eternal Sunshine of the Spotless
Mind.’” Screen Education 53 (2009): 139–144.
Academic Search Premier. Web. 3 Dec. 2011.
/

Endnotes

1 For this article, the circular plot of Eternal
Sunshine will be considered in plot, not story, order.

Williams.indd 49 2/13/13 5:02:56 PM

Copyright of Film Matters is the property of Intellect Ltd. and its content may not be copied or emailed to

multiple sites or posted to a listserv without the copyright holder’s express written permission. However, users

may print, download, or email articles for individual use.

Eternal Sunshine of the Spotless Mind? Mortality Salience, Death-Thought
Accessibility, and Self-Forgiveness

John M. McConnell

Wheaton College

ABSTRACT
Terror management theory claims the quintessential indicator of cultural adherence is
human self-esteem, and self-esteem is vital to suppressing death-thoughts into the uncon-
scious to buffer against existential fear. Guilt—an emotional response to cultural-based rule
violations—should therefore be an important motivation for self-forgiveness. In four studies,
mortality salience, death-thought accessibility, and self-forgiveness were negligibly related,
and offense severity, conciliatory behaviors, perceived forgiveness, and effort provided
negligible moderation. In Study 1, results did not support the initial interactional-moder-
ated-mediation model. In three follow-up studies, experimental-causal-chain analyses had
unsupportive findings. I provide implications for theory and research.

How happy is the blameless vestal’s lot!
The world forgetting, by the world forgot.
Eternal sunshine of the spotless mind!
Each pray’r accepted, and each wish resign’d
— Alexander Pope (1717/1796)

The depth and breadth of self-forgiveness literature has
lagged behind interpersonal forgiveness scholarship since
their commencement, but self-forgiveness has increased
in popularity among researchers more recently (Hall &
Fincham, 2005; McConnell, 2015). The number of self-
forgiveness publications has exponentially grown from
the early 1990s to present day (McConnell, 2015). Why,
how, and should people forgive themselves have now
been the subjects of partial and comprehensive
reviews, including a major edited volume (Davis,
Worthington, Hook, & Hill, 2013; Davis et al., 2015;
Fisher & Exline, 2010; Hall & Fincham, 2005; Lavelock,
Griffin, & Worthington, 2013; McConnell, 2015; Mullet,
Neto, & Rivi�ere, 2005; Shan & Xu, 2008; Tangney,
Boone, & Dearing, 2005; Webb, Hirsch, & Toussaint,
2011; Wohl & McLaughlin, 2014; Woodyatt,
Worthington, Wenzel, & Griffin, 2017; Worthington,
Witvliet, Pietrini, & Miller, 2007). However, historical
work in self-forgiveness and its modern wave of popular-
ity have mostly produced conceptual foundations
and atheoretical empirical findings. To encourage greater
theoretical specification as well as to warn against
mono-methodological, mono-measurement-type, and

mono-analytical biases, McConnell (2015) reviewed the
extant literature and proposed a conceptual-theoretical-
empirical (C-T-E) framework for self-forgiveness.
McConnell’s C-T-E framework organizes multiple phe-
nomena in a broad conceptual structure, provides rela-
tion specificity through theoretical foundations, and
packages scientific support in a coherent manner
(Fawcett, 1993). McConnell overviewed a multitude of
midrange theories to refine conceptual propositions and
empirical findings related to why, how, and should peo-
ple forgive themselves. In the current series of experi-
ments, I empirically tested if terror management theory
(TMT; Greenberg, Pyszczynski, & Solomon, 1986;
Pyszczynski, Greenberg, & Solomon, 1999) is an explan-
ation for why people forgive themselves as specified in
McConnell’s C-T-E framework. First, I broadly overview
conceptual and empirical scholarship in the origins of
self-forgiveness. Next, I specifically discuss self-forgive-
ness embedded within a TMT framework. Finally, I
describe four TMT self-forgiveness experiments along
with their implications for further theoretically
driven research.

Psychological origins of self-forgiveness

A substantial body of conceptual and empirical work
has highlighted the harmful personal effects of self-
unforgiveness (e.g., Davis et al., 2015; McConnell,

CONTACT John M. McConnell dr.john.m.mcconnell@gmail.com School of Psychology, Counseling, and Family Therapy, 501 College Avenue,
Wheaton, IL 60187-5501, USA.
� 2019 Taylor & Francis Group, LLC

BASIC AND APPLIED SOCIAL PSYCHOLOGY
2018, VOL. 40, NO. 6, 341–373
https://doi.org/10.1080/01973533.2018.1513361

http://crossmark.crossref.org/dialog/?doi=10.1080/01973533.2018.1513361&domain=pdf

http://orcid.org/0000-0002-4742-9854

http://www.tandfonline.com

2015). For this reason, scholars view self-forgiveness,
in part, as a utilitarian or adaptive process that allows
people to move beyond the harmful biological, psy-
chological, social, and existential/spiritual effects of
acute and chronic self-conscious emotions. Consistent
with Lazarus and Folkman’s (1984) stress and coping
model, self-forgiveness may be a fundamental well-
being safeguard (McConnell, 2015; Worthington et al.,
2007). Worthington and colleagues (2007) indicated
that self-forgiveness is both a problem-focused and
an emotion-focused coping strategy that allows
individuals to buffer against the unhealthy outcomes
of self-conscious emotions. In support, a large body of
correlational research has found that self-forgiveness
is positively related to physical, psychological, social,
and existential/spiritual well-being (see Davis et al.,
2015; McConnell, 2015). In Davis et al.’s (2015) meta-
analytic review of correlational research, they found
self-forgiveness related to both physical and psycho-
logical well-being. Of particular importance to the
current studies, Davis et al. found self-forgiveness was
related to life-satisfaction/meaning-in-life. This finding
also is consistent with qualitative research. For
example, people have sought self-forgiveness (Ferrell,
Otis-Green, Baird, & Garcia, 2014) and forgiveness
(Heflick, 2005) near the end of their lives.

Although it is reasonable to suggest self-forgiveness
is related to well-being in multiple domains of human
functioning based on the aforementioned findings,
atheoretical and nonexperimental methods do not
support that self-forgiveness causes increased well-
being, nor do they support that self-forgiveness has
existed to regulate well-being (Kline, 2015; Spencer,
Zanna, & Fong, 2005; Trafimow, 2015). In an attempt
to provide theoretical specification and increase
experimental rigor in this subfield, McConnell (2015)
indicated that TMT may be a cogent explanation for
the psychological origins of self-forgiveness and its
connection to (existential) well-being in correl-
ational research.

Terror management theory

According to TMT, avoiding thoughts of death moti-
vates people’s intrapsychic and social behavior because
the finite nature of human beings can be a difficult
reality to accept (Greenberg et al., 1986; Pyszczynski
et al., 1999). Death awareness—when people become
aware of their inevitable death—can be particularly
anxiety provoking because humans are biologically
ingrained with desires to sustain their lives. Yet all
people know they must eventually perish and strongly

desire that death would not be a fundamental human
condition. All human beings therefore have intrapsy-
chic conflicts, or existential anxiety, as they struggle to
reconcile their desire for immortality with the know-
ledge of inevitable death (Greenberg et al., 1986;
Pyszczynski et al., 1999). As Hayes, Schimel, Arndt,
and Faucher (2010) stated, “The knowledge of death,
coupled with a basic desire for continued life, creates
an existential dilemma capable of producing poten-
tially paralyzing anxiety, or terror” (p. 700).

According to TMT, people do not consciously
experience this intrapsychic conflict for most of their
waking life because humans have devised elaborate
ways to suppress death awareness (Greenberg et al.,
1986; Pyszczynski et al., 1999). For the times people
are reminded of death, humans have developed a dual
process component of proximal and distal defenses to
suppress death-thoughts into the unconscious
(Greenberg et al., 1986; Hayes et al., 2010; Pyszczynski
et al., 1999). Proximal defenses are rational responses
to death awareness that involve thought suppression,
distraction, or underestimation of current death vul-
nerability (Hayes et al., 2010). Upon reminders of
death, people initially use proximal defenses to sup-
press death awareness outside of their focal attention
into the unconscious (Hayes et al., 2010; Pyszczynski
et al., 1999). However, proximal defenses require sus-
tained cognitive resources, and over time unconscious
death-thoughts presumably become more readily
accessible just outside of conscious awareness (Hayes
et al., 2010; Pyszczynski et al., 1999), or “on the
fringes of consciousness—highly accessible but not in
current focal attention” (Pyszczynski, Solomon, &
Greenberg, 2015, p. 15). In response to this death-
thought suppression and delayed rebound effect
(Trafimow & Hughes, 2012), people use distal
defenses to further buffer against increasingly salient
preconscious levels of death-thoughts as proximal
defenses diminish (Hayes et al., 2010; Pyszczynski
et al., 1999).

The proponents of TMT believe that humans devel-
oped cultural worldviews and self-esteem as distal
defenses against death awareness. First, cultural world-
views buffer against existential anxiety by convincing
humans that there is order in a world of chaos, pro-
viding conscious and time-consuming activities, and
ensuring them that there is purpose and meaning to
their lives if they live up to certain expectations
(Greenberg et al., 1986; Pyszczynski et al., 1999).
In other words, culture provides humans illusory
senses of immediate biological safety, distraction, and
value. Second, self-esteem provides reassurance to

342 J. M. MCCONNELL

people that they are valuable and have lived up to
their cultural expectations (Greenberg et al., 1986;
Pyszczynski et al., 1999). According to TMT, cultural
worldviews and self-esteem work in combination to
convince humans that despite biological death they
will persist on with real (e.g., heaven, reincarnation)
or symbolic (e.g., fame, societal contribution) everlast-
ing life (Greenberg et al., 1986). If biological death is
not the end, then the terror it evokes has a dimin-
ished psychological impact consciously.

TMT has amassed an impressive amount of empir-
ical support over three decades with more than 500
studies to date (Darrell & Pyszczynski, 2016), includ-
ing multiple meta-analyses (Burke, Kosloff, & Landau,
2013; Burke, Martens, & Faucher, 2010; Martens,
Burke, Schimel, & Faucher, 2011; Steinman &
Updegraff, 2015; vanDellen, Campbell, Hoyle, &
Bradfield, 2011). Scholars elsewhere have extensively
reviewed TMT’s theoretical modifications and experi-
mental evidence (e.g., Burke et al., 2010; Hayes et al.,
2010; Pyszczynski, Greenberg, & Solomon, 1997;
Pyszczynski et al., 1999; Pyszczynski et al., 2015;
Solomon, Greenberg, & Pyszczynski, 2004; Steinman
& Updegraff, 2015; Vail et al., 2012). Despite TMT’s
remarkable number of studies that have provided
empirical support and its popularity in social psych-
ology, at times it has been contradicted or unsup-
ported by empirical evidence (e.g., Proulx & Heine,
2008; Trafimow & Hughes, 2012). TMT also has
received considerable criticism (e.g., Martin & van
den Bos, 2014; see also Psychological Inquiry special
issues: Pyszczynski et al., 1997 [Vol. 8, Issue 1], and
Martin & Erber, 2006 [Vol. 17, Issue 4]). For instance,
two meta-analyses suggested that TMT’s tenets are
not culturally universal (Yen & Cheng, 2010) and
research effect sizes have been partially a by-product
of experimental bias (Yen & Cheng, 2013).

Although Pyszczynski et al. (1997) claimed that
TMT is a parsimonious theory to explain virtually all
human behavior, and that it explains the findings of
diverse theories of social motivation, other scholars
have made equal attempts with competing theories to
explain the social-cognitive phenomena and findings
in the purview of TMT (e.g., Heine, Proulx, & Vohs,
2006; Proulx, Inzlicht, & Harmon-Jones, 2012; see also
Martin & Erber, 2006 [Psychological Inquiry, Vol. 17,
Issue 4], and Proulx, 2012 [Social Cognition, Vol. 30,
Issue 6], for special issues). However, some evidence
has indicated that death may be a unique psycho-
logical threat. For example, Martens et al.’s (2011)
meta-analysis found that death awareness elicited
higher levels of defensiveness compared to meaning

or certainty threats when measurement occurred after
delays from the experimental manipulation. Recently,
scholars have attempted to integrate many theories of
psychological defense, including TMT, into one coher-
ent theoretical framework (e.g., George & Park, 2016;
Hart, 2014), or worked to go beyond integration to
unification in psychological theory (e.g., Henriques,
2011; Trafimow, 2012b; see also Gaj, 2016). Despite
these attempts, TMT currently remains a mainstay in
social psychology, and researchers have continued to
investigate three main hypotheses derived from the
theory: the mortality-salience (TMT Hy1), death-
thought accessibility (TMT Hy2), and anxiety-buffer
(TMT Hy3) hypotheses (Darrell & Pyszczynski, 2016;
Pyszczynski et al., 2015).

Mortality salience hypothesis
The TMT Hy1 posits that increased death awareness—
operationalized as mortality salience (MS)—will cause
people to cognitively suppress their death-related
thoughts, defend their own unique cultural world-
views, and bolster their self-esteem to buffer against
conscious awareness of death (Pyszczynski et al., 1999;
Pyszczynski et al., 2015). This hypothesis assumes that
MS will initially increase proximal defenses and even-
tually increase distal defenses once proximal ones
diminish. In other words, people will display increased
amounts of in-group identification, out-group dispar-
agement, and self-esteem enhancement due to MS
as cognitive suppression diminishes over time.
For instance, people have initially suppressed thoughts
of death following death awareness (Arndt, Cook,
Goldenberg, & Cox, 2007), but after a delay
distributed excessive hot sauce to other cultural group
members’ food (McGregor et al., 1998) or displayed
self-serving attributions (Mikulincer & Florian, 2002).
McConnell (2015) indicated that people may engage
in the process of self-forgiveness following MS as
a method to reaffirm cultural values and repair
self-esteem. If TMT is an explanatory framework for
self-forgiveness, then MS would lead to increased self-
forgiveness after a delay. No known work has tested
the MS hypothesis in relation to self-forgiveness.

Burke et al.’s (2010) meta-analysis of 277 experi-
ments found that increased initial MS had a so-called
moderate effect on increased distal defense behaviors.
Neutral (e.g., television) or negative (e.g., dental pain)
comparison conditions and gender negligibly moder-
ated the effect MS had on distal defenses, but being a
United States citizen or an undergraduate increased
the MS effect. When Yen and Cheng (2013) reana-
lyzed the literature in another meta-analysis,

BASIC AND APPLIED SOCIAL PSYCHOLOGY 343

experimental team moderated the MS effect as well.
The MS effect also increased with longer delays fol-
lowing initial reminders of death (Burke et al., 2010).
Time lapse from initial MS is an important theoretical
concept in TMT. Distal defenses are supposedly not
activated without a sufficient delay for increasingly
salient nonconscious death-thoughts to become
accessible (cf. Burke et al., 2010; Hayes et al., 2010;
Pyszczynski et al., 1999; Steinman & Updegraff, 2015;
Trafimow & Hughes, 2012).

Death-thought accessibility hypothesis
The TMT Hy2 posits thoughts of death just outside of
focal awareness—operationalized as death-thought
accessibility (DTA)—mediate between MS and distal
defenses (Hayes et al., 2010). In other words, TMT
theorists believe that initial MS has an effect on distal
defenses through DTA. Therefore, if TMT is the the-
oretical foundation for self-forgiveness, then initial
MS would lead to increased self-forgiveness after a
delay as DTA becomes more salient. No known work
has tested the DTA hypothesis in relation to self-
forgiveness.

A growing body of literature has supported that MS,
death-reminders, and other threats (e.g., cultural value,
self-esteem) lead to increased DTA following a delay.
Steinman and Updegraff’s (2015) meta-analysis of 99
experiments found a so-called moderate effect size on
DTA following existentially related threats, and more dis-
tractor tasks or longer delays increased the effect.
Neutral or negative comparison conditions and the
measurement type of DTA (e.g., word-fragment comple-
tion, reaction time) negligibly moderated the effect.
Despite the large body of evidence for the DTA hypoth-
esis, there also has been conflicting evidence. For
instance, in a series of six experiments, Trafimow and
Hughes (2012) found the exact opposite of TMT Hy2.
DTA actually decreased with delays. Similarly, other
researchers have found that DTA negligibly changed
with experimental manipulations in multiple unpublished
studies (e.g., Hart, 2014; Hughes, as cited in Hart, 2014).

Given TMT researchers believe DTA mediates exist-
ential threats and distal defenses, it is surprising at first
glance that there is a paucity of studies within TMT
research that directly test DTA as a mediator (Hayes
et al., 2010). Researchers have begun such work, and
multiple studies have implied that DTA either mediated
or moderated existential threat and defense (e.g., Cox,
Eaton, Ekas, & Van Enkevort, 2015; Cox, Reid-Arndt,
Arndt, & Moser, 2012; Das, Bushman, Bezemer,
Kerkhof, & Vermeulen, 2009; Echebarria-Echabe, 2013;
Fransen, Fennis, Pruyn, & Das, 2008; Hayes et al., 2015;

Motyl et al., 2013; Vail, Arndt, Motyl, & Pyszczynski,
2012). Yet there are inherent challenges in demonstrat-
ing DTA mediation for three main reasons. First, meas-
urement-of-mediation designs cannot rationally provide
direct support of mediation (Kline, 2015; Spencer et al.,
2005; Tate, 2015; see Trafimow, 2015 [Basic and
Applied Social Psychology, Vol. 37, Issue 4], for a special
issue on mediation). Second, pure experimental-causal-
chain designs are difficult to establish in TMT because
manipulating implicit processes such as DTA are diffi-
cult, if not impossible, without evoking MS or anxiety-
buffer threats (Hayes et al., 2010; Hayes & Schimel,
2018). Third, a measurement-of-mediation design may
inadvertently expose a comparison group to death-
related stimuli while filling in a DTA measure.
Recently, Hayes and Schimel (2018) demonstrated that
the measurement of DTA in comparison groups obfus-
cated MS threat effects on distal defenses, but there was
a rebound effect after a post-DTA measurement delay.
Hayes and Schimel (2018) indicated that DTA measure-
ment itself becomes a MS threat, and therefore meas-
urement artifacts minimize the chances of empirically
demonstrating the mediational aspect of the DTA
hypothesis in one experiment.

Anxiety-buffer hypothesis
The TMT Hy3 posits strengthening self-esteem pro-
tects against anxiety with subsequent existential
threats. Inversely, weakening self-esteem makes people
vulnerable to existential anxiety (Hayes et al., 2010;
Pyszczynski, Greenberg, Solomon, Arndt, & Schimel,
2004; Pyszczynski et al., 1999, 2015). According to the
proponents of TMT, the anxiety-buffer regulates exist-
ential anxiety, DTA, and worldview defenses with and
without explicit reminders of death. Thus, if TMT is a
theoretical underpinning of self-forgiveness, then MS
and subsequent DTA would lead to self-forgiveness
when self-esteem is low. In addition, even in the
absence of MS threats, decreased self-esteem would
lead to increased DTA and, in turn, increased self-for-
giveness. No known work has tested the anxiety-buffer
hypothesis in relation to self-forgiveness.

A large body of literature has supported that self-
esteem serves as an anxiety buffer against MS, world-
view threats, DTA, and other existentially related phe-
nomena (Pyszczynski et al., 2004; Hayes et al., 2010;
Schimel, Landau, & Hayes, 2008; Steinman &
Updegraff, 2015). Research has supported that self-
esteem serves as a buffer against general and existential
anxiety and that people bolster their self-esteem to pro-
tect against existential angst (Pyszczynski et al., 2004;
Schimel et al., 2008). For instance, Greenberg et al.

344 J. M. MCCONNELL

(1992) found increased self-esteem through positive per-
sonality feedback decreased anxiety about a death-
related video.

Hypotheses summary
TMT Hy1, TMT Hy2, and TMT Hy3 each represent a
unique facet of TMT’s theoretical underpinnings and
empirical research. Hayes et al. (2010) offered an illus-
trative theoretical process model of TMT that depicts
the temporal sequence these hypotheses claim. In
short, MS leads to activation of the anxiety buffer
either immediately or following a delay if proximal
defenses are initially active. If the anxiety buffer is
weak, then people will experience high levels of DTA
and subsequent distal defenses to decrease DTA. If
the anxiety buffer is strong, then people will experi-
ence low levels of DTA (Hayes et al., 2010). Interested
readers can turn to Hayes et al. (2010) for a compre-
hensive review and description of their TMT process
model, including various nuance processes. Although
researchers often have isolated these hypotheses, the
processes they describe are inextricably interrelated.
Therefore, I tested all three hypotheses in the current
series of studies to comprehensively evaluate if TMT
is an explanatory framework for self-forgiveness.

Self-forgiveness: A self-esteem distal
defense process?

According to TMT, self-esteem is a fundamental psy-
chological mechanism that buffers against existential
anxiety by convincing people they are worthy of either
real or symbolic eternal life because they have lived
up to their cultural standards (Greenberg et al., 1986;
Pyszczynski et al., 1999, 2004). People can fail to meet
cultural expectations in many ways, and modern lexi-
con includes numerous words to describe such wrong-
doings. For instance, people can commit wrongdoings
that are immoral (e.g., wicked deed, evil doing, atro-
city), unethical (e.g., violation, infraction, misconduct),
illegal (e.g., offense, crime, misdemeanor, felony),
interpersonal (e.g., let down, betrayal, harm, hurt,
wound, injure), self-inflicted (e.g., self-harm, self-
injury, self-handicap, mistake, blunder), and religious/
spiritual (e.g., sin, transgression, bad karma, p�apa,
dhanb) in nature. McConnell (2015) theorized that
guilt and shame for cultural worldview violations were
the nexus among self-esteem, death anxiety, and exist-
ential motivations for self-forgiveness. If Solomon,
Greenberg, and Pyszczynski’s (1997, p. 70) fundamen-
tal proposition that death is central to “virtually every-
thing that we think, feel, say, and do” is true, then it

would follow: (a) culture-based rule violations cause
guilt and shame, (b) which reduce self-esteem because
of their death-related implications, and (c) self-for-
giveness is an attempt to mollify existential angst
based in such self-conscious emotions.

To expand on this theoretical extrapolation of
TMT, people experience guilt or shame when they
have violated cultural rules and they have taken
responsibility for their actions (McConnell, 2015).
Guilt and shame are self-conscious signifiers to indi-
viduals that they have violated cultural expectations
required for real or symbolic eternal life. As a result,
the self-conscious nature of guilt and shame reduce
self-esteem. With the anxiety buffer weakened, MS
and DTA will cause people to experience existential
anxiety as they struggle to reconcile their finite lives
and imperfect character. Existential anxiety will be
amplified to the extent that they consciously and
unconsciously believe cultural norm violators are
unworthy to inherent real or symbolic everlasting life.
In this context, death has frightening implications
when self-esteem does not provide reassurance that
biological death will simply be a new existential awak-
ening. Faced with this existential anxiety, people are
compelled to repair their self-esteem and existential
well-being through self-forgiveness. Overall, if the
anxiety buffer is weak (TMT Hy3), MS should lead to
DTA after a delay (TMT Hy2), and in turn the distal
defense of self-forgiveness (TMT Hy1).

If these assumptions were true, then TMT would
shed an interesting light on why guilt, shame, and
decreased self-forgiveness have consistently related to
depression, anxiety, and decreased overall well-being
(e.g., Davis et al., 2015; Kim, Thibodeau, & Jorgensen,
2011; McConnell, 2015; Tangney & Dearing, 2002)
and why death anxiety has played a unique role in
psychopathology (Iverach, Menzies, & Menzies, 2014).
Several studies have indicated that people experienced
guilt following MS (e.g., Arndt, Greenberg, Solomon,
Pyszczynski, & Schimel, 1999; Fergus & Valentiner,
2012; Goldenberg, Heflick, & Cooper, 2008; Harrison
& Mallet, 2013; Hirschberger, Florian, & Mikulincer,
2002), and one qualitative study revealed that some
TMT participants wrote about guilt during MS
manipulations (Kastenbaum & Heflick, 2011). In add-
ition, anecdotal evidence indicated that some people
have sought self-forgiveness (Ferrell et al., 2014)
and forgiveness (Heflick, 2005) near the end of their
lives. Researchers have studied forgiveness and recon-
ciliation in the context of TMT (Anglin, 2014;
Schimel, Wohl, & Williams, 2006; Van Tongeren,
Green, Davis, Worthington, & Reid, 2013; Wilson &

BASIC AND APPLIED SOCIAL PSYCHOLOGY 345

Bernas, 2011), but the current studies are the first
known attempts to evaluate if TMT is an explanatory
framework for self-forgiveness.

General method

In a series of four experiments, I examined the valid-
ity of TMT as a theoretical foundation for self-forgive-
ness by systematically testing TMT Hy1, TMT Hy2,
and TMT Hy3. I expected that MS would lead to the
distal defense of self-forgiveness (TMT Hy1) through
DTA after a delay (TMT Hy2) in the context of a
weakened anxiety buffer from recalling an interper-
sonal offense (TMT Hy3). I considered TMT’s meth-
odological designs, procedures, and measurements
known to elicit the highest effect sizes (cf. Burke
et al., 2010, 2013; Hayes et al., 2010; Martens et al.,
2011; Steinman & Updegraff, 2015; Yen & Cheng,
2010, 2013) in the conceptualization of the current
studies, although they were designed and conducted
before the online publication of Hayes and Schimel’s
(2018) study. The following is an overview of an ini-
tial attempt to test all three of TMT’s experimental
hypotheses in one study and an experimental-causal-
chain procedure designed to test the sequence in a
step-by-step manner across three follow-up studies.
All studies received prior approval from the author’s
Institutional Research Review Boards.

Experiment

1

In Experiment 1, I tested an interactional-moderated-
mediation model among MS, DTA, and self-
forgiveness. Specifically, following an anxiety-buffer
threat of a recalled recent interpersonal offense, I
tested if DTA mediated the relation between MS and
state self-forgiveness after a delay, and if this relation
was moderated by offense severity, conciliatory behav-
iors, perceived forgiveness, or effort. I selected these
four moderators to control for their statistical variance
due to their consistent relation to the process of self-
forgiveness. To maximize construct validity, I utilized
state measures of guilt and self-forgiveness to create
the latent outcome variable “self-forgiveness” and
assessed participants’ level of responsibility to rule out
pseudo self-forgiveness. I excluded shame because of
its inconsistent relation to self-forgiveness (see
McConnell, 2015, for a full rationale of contextual
variables and construct validity). I used a randomized,
post-test-only, double-blind, comparison-group design
(MS vs. dental pain salience [DP]) combined with a
measurement-of-mediation procedure (Shadish, Cook,

& Campbell, 2002; Spencer et al., 2005). I tested the
following hypotheses in a single study:

1. Mediation hypothesis. Mediation can be minim-
ally alluded to when an experimentally manipu-
lated independent/predictor variable (MS) affects
a dependent/criterion variable (self-forgiveness)
but only through a measured third variable
(DTA). There also must be a conceptually based
temporal sequence (Kline, 2015; Spencer et al.,
2005; Tate, 2015). Partial mediation is suggested
when the relation between the predictor and cri-
terion variables is reduced in part after account-
ing for the mediating effect, whereas full
mediation is implied when the relation is nullified
(Baron & Kenny, 1986; Kline, 2015; Preacher,
Rucker, & Hayes, 2007; Sobel, 1982). Because
TMT theorists posit MS produces distal defenses
through emerging preconscious death-thoughts
(e.g., Hayes et al., 2010), I predicted that DTA
(TMT Hy2) would fully mediate between MS and
self-forgiveness (TMT Hy1) following an anxiety-
buffer threat of recalling an interpersonal offense
(TMT Hy3).

2. Moderated-mediation hypothesis. A moderated
mediation is suggested when the strength or dir-
ection of a mediated relation is contingent on one
or more ancillary variables (Preacher et al., 2007).
In other words, the presence of other variables
strengthens or weakens the relation of the medi-
ated process. MS and DTA may help explain why
people strive for self-forgiveness, although offense
severity, conciliatory behaviors, perceived forgive-
ness, and effort to reduce emotions play import-
ant roles in how easily people accomplish self-
forgiveness. Therefore, I predicted that higher lev-
els of conciliatory behaviors, perceived forgive-
ness, and effort would increase the relation
between DTA and self-forgiveness, whereas higher
levels of offense severity would decrease the rela-
tion (McConnell, 2015).

3. Interactional-moderated-mediation hypothesis.
Interactional-moderated-mediation is implied
when a moderated-mediation relation is not
apparent at all levels of the independent variable
(Hayes et al., 2010; Preacher et al., 2007). In other
words, the experimental group would exhibit a
moderated-mediation relation, but the compari-
son group would not. Participants in a TMT com-
parison group are less likely to have sufficient
DTA to exhibit any distal defenses because they
are not explicitly reminded of death, although

346 J. M. MCCONNELL

they may have inadvertent implicit exposure
(Burke et al., 2010; Hayes et al., 2010; Hayes &
Schimel, 2018; Steinman & Updegraff, 2015).
Therefore, I predicted that the MS experimental
group would have more DTA and self-forgiveness
than the DP comparison group, and in turn DTA
would positively relate to self-forgiveness for the
MS group but not for the DP group.

Method

Participants
Participants solicited from introductory to psychology
(n ¼ 175; 77.8%) or interpersonal relations (n ¼ 50;
22.2%) classes at a public university were recruited via
a posting on the psychology department’s website and
were allowed to participate if they were at least
18 years old. An initial 226 participants participated;
however, one person was not included in the analyses
because he admitted to random responding. See
Table 1 for a summary of demographic characteristics.

Procedure
Participants registered for the experiment using an
online scheduling program. Similar to previous
research, potential participants were recruited to
enroll in the study if they felt poorly about hurting
one person close to them with their words or actions
within the last month (Exline et al., 2007, 2011; Fisher
& Exline, 2006; Hall & Fincham, 2008). Upon arriving
at an experimental room within 3 days after recruit-
ment (Hall & Fincham, 2008), participants were
greeted by my research assistant who was blind to the
experimental conditions. Participants first completed
an informed consent and then an offense packet while
seated privately in a classroom setting. The first task
in the offense packet was to recall and briefly write
about their recent interpersonal wrongdoing in order
to induce an anxiety-buffer threat. Participants then
filled in counterbalanced measures of responsibility,
offense severity, conciliatory behavior, and perceived
forgiveness. (See Figure 1 for a visual depiction of the
experimental sequence and the Materials section for a
description of packets used in Experiment 1.)

Participants then completed the death awareness
packet, which consisted of several personality meas-
ures administered under the guise of collecting data
for another researcher who “asked us to collect data
for them” as other TMT researchers have done (see
Burke et al., 2010). I included this deception to reduce
suspicion about filling in the criterion and effort
measures after the experimental manipulation rather

than directly after the offense recall (Burke et al.,
2010; Hayes et al., 2010). Embedded at the end of this
packet, I randomly assigned participants to one of two
conditions blind to the research assistant. Participants
were asked to respond to two open-ended questions
designed to increase either (a) MS (n ¼ 112; 49.8%) or
(b) DP (n ¼ 113; 50.2%; Rosenblatt, Greenberg,
Solomon, Pyszczynski, & Lyon, 1989). My research
assistant then gave them a number of additional coun-
terbalanced measures to delay time after the experi-
mental manipulation. This distractor-delay aimed to
increase the MS effect on DTA and dependent meas-
ures, as well as rule out effects of social desirability
and affect (cf. Burke et al., 2010; Hayes et al., 2010;
Steinman & Updegraff, 2015; Trafimow &
Hughes, 2012).

Next, both the experimental (MS) and comparison
(DP) groups filled in a DTA word-fragment comple-
tion task. The research assistant then told participants
that he forgot to administer some measures from the
prior study. All participants then filled in state self-
forgiveness and guilt scales in counterbalanced orders
followed by an effort to reduce emotions measure.
Participants subsequently completed a demographic
questionnaire. Next, I personally interviewed partici-
pants about any emotional distress, demand character-
istics, and hypotheses guessing. Participants reported
no adverse events. Interviews revealed zero partici-
pants, including those who were suspicious of poten-
tial deception (n ¼ 54), were able to guess the
rationale for the study’s methodology or its hypothe-
ses, and no one was even aware of TMT itself. Finally,
I debriefed them about the deception and gave them
research credit through an online database.

Materials
I used the following offense and death awareness
packets, as well as the distractor-delay, DTA, and cri-
terion measures. See Table 2 for Experiment 1
scale properties.

Offense packet. The following offense recall instruc-
tions and counterbalanced measures were given to
participants to assess their wrongdoing and related
social-cognitive phenomena.

Offense recall. Participants received the following
instructions:

Please recall a recent event in the past month in
which you hurt someone close to you by something
you said or did. Recall an event in which you
currently—right now—feel blameworthy and

BASIC AND APPLIED SOCIAL PSYCHOLOGY 347

Ta
b
le

1.
D
em

og
ra
p
h
ic
ch
ar
ac
te
ri
st
ic
s

fo
r
Ex
p
er
im
en
ts

1–
4.

M
(S
D
)

Ra
n
g
e

1
2

3
4

1
2
3
4

A
g
e

19

.7
7

(1
.8
1)

39

.5
6

(1
5.
42
)

42
.5
7
(1
6.
19
)

39

.6
8

(1

5.
67

)

18
/3
0

18
/8
0

18
/8
8

18
/8
4

Ed
uc
at
io
n

13

.1
2

(1
.3
8)

13
.6
2
(3
.2
5)

13

.7
8

(3
.6
3)

13
.7
2
(3
.5
8)

12
/2
1

0/
25

0/
30

0/
30

D
ay
s
si
n
ce

of
fe
n
se

17

.3
8

(1
0.
30
)

98

.5
3

(1
11
.1
6)

–
87
.6
2
(1
05
.1
1)

1/
31

1/
36
5

–
1/
36
5

Ti
m
es

co
m

m
it
te
d

2.
93

(8
.8
8)

15
78

.2
4

(3
37
11
.2
9)

–
13
31

.4
1

(3
36
50
.3
3)

1/
10
0

1/
99
99
99

–
1/
10
00
00
0

G
en
d
er

Re
lig
io
n

1
2
3
4
1
2
3
4

M
an

79
(3
5.
1%

)
38
1
(4
0.
9%

)
16
8
(3
5.
3%

)
38
3
(4
2.
9%

)
C
at
h
ol
ic

55
(2
4.
4%

)
16
8
(1
8.
0%

)
86

(1
8.
1%

)
15
8
(1
7.
7%

)
W
om

an
14
5
(6
4.
4%

)
55
0
(5
9.
0%

)
30
6
(6
4.
3%

)
50
7
(5
6.
8%

)
Pr
ot
es
ta
n
t

86
(3
8.
2%

)
35
5
(3
8.
1%

)
16
6
(3
4.
8%

)
33
7
(3
7.
8%

)
O
th
er

1
(0
.4
%
)

1
(0
.1
%
)

2
(0
.4
%
)

2
(0
.4
%
)

O
rt
h
od

ox
0
(0
%
)

7
(0
.8
%
)

2
(0
.4
%
)

5
(0
.6
%
)

Je
w
is
h

5
(2
.2
%
)

16
(1
.7
%
)

8
(1
.7
%
)

19
(2
.1
%
)

M
us
lim

3
(1
.3
%
)

6
(0
.6
%
)

2
(0
.4
%
)

4
(0
.4
%
)

Bu
d
d
h
is
t

3
(1
.3
%
)

15
(1
.6
%
)

9
(1
.9
%
)

15
(1
.7
%
)

H
in
d
u

1
(0
.4
%
)
4
(0
.4
%
)

6
(1
.3
%
)

4
(0
.4
%
)

Et
h
n
ic
it
y

A
g
n
os
ti
c

25
(1
1.
1%

)
65

(7
.0
%
)

33
(6
.9
%
)

49
(5
.5
%
)

1
2
3
4

A
th
ei
st

20
(8
.9
%
)

55
(5
.9
%
)

19
(4
.0
%
)

53
(5
.9
%
)

W
h
it
e

18
4
(8
1.
8%

)
72
0
(7
7.
3%

)
38
0
(7
9.
8%

)
67
9
(7
6.
1%

)
N
on

e
27

(1
2%

)
24
1
(2
5.
9%

)
14
5
(3
0.
4%

)
24
8
(2
7.
8%

)
A
fr
ic
an

A
m
er
ic
an

/B
la
ck

20
(8
.9
%
)

78
(8
.4
%
)

36
(7
.6
%
)

85
(9
.9
%
)

La
ti
n
o/
a

2
(0
.9
%
)

67
(7
.2
%
)

18
(3
.8
%
)

46
(5
.2
%
)

M
ar
it
al
St
at
us

A
si
an

or
A
si
an

A
m
er
ic
an
5
(2
.2
%
)

33
(3
.5
%
)

22
(4
.7
%
)

38
(4
.2
%
)

1
2
3
4

N
at
iv
e
A
m
er
ic
an

0
(0
.0
%
)

6
(0
.6
%
)

7
(1
.5
%
)

11
(1
.2
%
)

Si
n
g
le

20
3
(9
0.
2%

)
27
2
(2
9.
2%

)
11
7
(2
4.
6%

)
28
7
(3
2.
2%

)
Pa
ci
fic

Is
la
n
d
er

0
(0
.0
%
)
1
(0
.1
%
)
2
(0
.4
%
)
4
(0
.4
%
)

C
oh

ab
it
at
in
g

16
(7
.1
%
)

15
9
(1
7.
1%

)
49

(1
0.
3%

)
13
6
(1
5.
2%

)
N
.
A
fr
ic
an

an
d

M
.
Ea
st
er
n

0
(0
.0
%
)
1
(0
.1
%
)
2
(0
.4
%
)
4
(0
.4
%
)

M
ar
ri
ed

4
(1
.8
%
)

36
5
(3
9.
2%

)
23
0
(4
8.
3%

)
33
7
(3
7.
8%

)
Bi
ra
ci
al

8
(3
.6
%
)

15
(1
.6
%
)
2
(0
.4
%
)

6
(0
.7
%
)

Se
p
ar
at
ed

2
(0
.9
%
)
16
(1
.7
%
)

11
(2
.3
%
)

28
(3
.1
%
)

M
ul
ti
ra
ci
al

5
(2
.2
%
)
7
(0
.8
%
)
2
(0
.4
%
)

12
(1
.3
5%

)
D
iv
or
ce
d

0
(0
.0
%
)

91
(9
.8
%
)

45
(9
.5
%
)

76
(8
.5
%
)

U
n
sp
ec
ifi
ed

1
(0
.4
%
)
4
(0
.4
%
)

5
(1
.1
%
)

4
(0
.4
%
)

W
id
ow

ed
0
(0
.0
%
)

29
(3
.1
%
)

24
(5
.0
%
)

28
(3
.1
%
)

N
ot
e.

Ex
p
er
im
en
ta
l
co
n
d
it
io
n
s
co
m
b
in
ed

w
it
h
in

st
ud

ie
s,

1:

n
¼
22
5;

2:
n
¼
93
2;

3:
n
¼
47
6;

4:
n
¼
89
2.

348 J. M. MCCONNELL

personally upset about what you did. For a moment,
visualize in your mind the events and the interactions
you may have had with the person you offended. Try
to visualize the person and recall what happened.
Now please briefly describe the event in five to
seven sentences.

Forgiveness and self-forgiveness researchers have
used similar recalls in previous research (Exline et al.,
2007; Fisher & Exline, 2006; Hall & Fincham, 2008;
McCullough et al., 1998). My research assistant and I
reviewed offense recalls to ensure that participants
adequately responded to the prompt with an interper-
sonal offense.

Responsibility. The Revised Causal Dimension Scale
(CDSII; McAuley, Duncan, & Russell, 1992) consists
of four separate subscales—Causality, External
Control, Personal Control, and Stability—which all
have demonstrated adequate levels of internal consist-
ency reliabilities in a previous sample. All nine items
are rated with a 9-point bipolar-type scale (e.g., Is the
cause(s) something: permanent vs. temporary; inside
of you vs. outside of you). The CDSII also has dem-
onstrated construct validity as evidenced by content
validity and confirmatory factor analysis (McAuley
et al., 1992). Although the internal consistencies for
the external control (a ¼ .73; M ¼ 11.34, SD ¼ 5.92)
and personal control (a ¼ .78; M ¼ 20.78, SD ¼ 5.56)
were adequate, they were not for causality (a ¼ .58;
M ¼ 16.51, SD ¼ 5.46) and stability (a ¼ .35;
M ¼ 11.67, SD ¼ 4.93) in Experiment 1.

Perceived offense severity. The Perceived Offense
Severity Scale (POS; Hall & Fincham, 2008) is a three-

item measure on a 7-point Likert-type scale, from 1
(very positively) to 7 (very negatively), that assesses how
people believe they affected the victim, themselves, and
their relationship with the victim (e.g., How did your
behavior affect the victim?). The scale has demonstrated
an adequate level of internal consistency reliability in a
previous sample (Hall & Fincham, 2008). Hall and
Fincham (2008) found the POS had a negative correl-
ation with self-forgiveness, which supported its conver-
gent validity (Hall & Fincham, 2008). The internal
consistency was less than adequate (a ¼ .53; M¼ 17.25,
SD ¼ 2.53) in Experiment 1.

Conciliatory behaviors. A modified version of the CBS
(McCullough et al., 1997) developed by McConnell,
Dixon, and Finch (2012) is a five-item measure (e.g., I
tried/will try to make amends or compensations) rated
on a 7-point Likert-type scale from 1 (strongly dis-
agree) to 7 (strongly agree). The modified CBS demon-
strated a desirable level of internal consistency
reliability in a previous sample (McConnell et al.,
2012). The modified CBS has good content validity
because it includes all components of an effective
apology (Olshtain, 1989; Weiner et al., 1991). The
modified CBS also has demonstrated construct validity
as evidenced by confirmatory factor analysis
(McConnell et al., 2012). McConnell et al. (2012)
found that the CBS had a positive correlation with
perceived forgiveness, which supported its convergent
validity. The internal consistency was adequate
(a ¼ .85; M ¼ 27.28, SD ¼ 7.75) in Experiment 1.

Perceived forgiveness. To quantify participants’ levels
of perceived forgiveness, I used two items developed

Additional MaterialsOffense Packet

Offense Recall

Counterbalanced

Responsibility
(CDS-II)

Perceived Offense
Severity
(POS)

Conciliatory
Behaviors

(CBS)

Perceived
Forgiveness
(PF-Victim)

Death Awareness
Packet

Distractor-Delay
Measures

Mediating Measure Criterion Measures

Cover
Story

Narcissism
(NPI-40)

Personality
(BFI-44)

M
O
R
T
A
L
I
T
Y

S
A
L
I
E
N
C
E

Counterbalanced

Social
Desirability
(M-C SDS)

Mood
(PANAS)

Puzzle Task

Death-Thought
Accessibility
(DTA Word
Completion

Task)

Counterbalanced

Self-Forgiveness
(SSFS)

Guilt
(SSGS)

Effort
(ES)

Demographic
Questionnaire

D
E
N
T
A
L

P
A
I
N

Counterbalanced
Social
Desirability
(M-C SDS)
Mood
(PANAS)
Puzzle Task

Death-Thought
Accessibility
(DTA Word
Completion

Task)
Counterbalanced
Self-Forgiveness
(SSFS)
Guilt
(SSGS)
Effort
(ES)
Demographic
Questionnaire

Threat
Condition

(MAPS)

Mortality

Salience

OR

Dental
Pain

Salience

D
E
C
E
P
T
I
O
N

Figure 1. Graphical representation of the experimental procedure for Experiment 1.

BASIC AND APPLIED SOCIAL PSYCHOLOGY 349

Ta
b
le

2.
Sc
al
e

p
ro
p
er
ti
es

fo
r
Ex
p
er
im
en
ts
1–
4.

a
M

SD
Ra
n
g
e

VI
F

Sk
ew

n
es
s

Ku
rt
os
is

1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4

D
TA

†�
–

–
–

–
1.
66

–
1.
88

1.
79

1.
09

–
0.
99

1.
02

0/
5

–
0/
5

0/
6

1.
12

–
1.
07

1.
05

0.
39

–
�0

.8
1

0.
27

�0

.3
4

–
�0
.3
4

0.
26

SF
FA

.8
7

.8
7

–

.8
6

19

.0
6

21

.3
5

–
21
.7
2

5.
75

5.
72

–
5.
66

8/
32

8/
32

–
8/
32

3.
15

2.
83

–
3.
09

0.
16

�0

.1
9

–
�0

.2
0

�0

.5
0

�0

.4
2

–
�0

.3
1

SF
B

.9
0

.9
1

–

.9
2

26
.8
0

27
.7
8

–
28

.0
0

6.
18

6.
56

–
6.
59

10
/3
6

9/
36

–
9/
36

2.
69

2.
73

–
3.
33

�0
.5
6

�0

.7
9

–
�0
.9
0

�0
.3
8

0.
07

–
0.
36

H
FS

–
–

.7
1

–
–

–
28
.6
3

–
–

–
6.
63

–
–

–
6/
42

–
–

–
2.
02

–
–

–
0.
00

–
–
–
�0

.2
7

–
PP
RS

–
.8
1

–
.8
0

–
36

.6
7

–
36

.5
4

–
11

.2
5

–
11
.2
4

–
5/
50

–
5/
50

–
1.
28

–
1.
30

–
�0
.5
4
–
�0
.5
6
–
�0

.4
8

–
�0
.4
2

C
D
SI
I:
Ex
te
rn
al

.7
3

–
–

–
11
.3
4

–
–

–
5.
92

–
–

–
3/
27

–
–

–
1.
44

–
–

–
0.
40

–
–
–
�0
.6
7
–
–

–
C
D
SI
I:
Pe
rs
on

al
.7
8

–
–

–
20
.7
8

–
–

–
5.
56

–
–
–
3/
27
–
–

–
1.
39

–
–

–
�1

.0
2

–
–

–
0.
65

–
–

–
PO

S
.5
3

.6
6

–

.7
5

17
.2
5

15

.2
6

–
15

.0
7

2.
53

3.
36

–
3.
61

5/
21

3/
21

–
3/
21

1.
50

1.
43

–

1.
41

�0
.8
7

�0
.2
4

–
�0

.3
2

2.
01

0.
21

–
0.
28

SS
G
S:

G
ui
lt

(R
)

.7
9
.8
7

–

.8
9

11
.6
8

16
.2
4

–
16
.0
2

4.
33

5.
84

–
5.
89

5/
25

5/
25

–
5/
25

1.
99

1.
81

–
1.
97

0.
73

�0

.3
6

–
�0
.2
4

0.
53

�0
.9
2

–
�1
.0
2

C
BS

.8
5

–
–

–
27

.2
8

–
–

–
7.
75

–
–

–
5/
35

–
–

–
1.
58

–
–
–
�0

.9
4

–
–
–
�0

.1
7

–
–

–
PF
-V
ic
ti
m
†

–
–
–
–

3.
41

–
–

–
1.
27

–
–

–
1/
5

–
–

–
1.
59

–
–
–
�0

.4
6

–
–
–
�0

.7
4

–
–

–
ES

.7
5
–
–

–
10
.4
3

–
–

–
5.
09

–
–

–
0/
20

–
–

–
1.
35

–
–
–
�0

.2
9

–
–
–
�0
.6
8
–
–

–
RS
ES

–
.8
8

.9
0
.8
9

–
18
.4
9

19
.7
5

18
.2
8

–
6.
27

6.
25

6.
55

–
0/
30

0/
30
0/
30

–
2.
09

2.
31

2.
44

–
�0

.0
1

�0
.2
0

�0
.1
7

–
�0

.5
2

�0
.4
8

�0
.3
5

M
-C

SD
S(
K
R
)

.7
7
.8
1
.7
8
.7
9

13
.7
1

17

.0
8

17
.9
0

16
.7
5

5.
02

5.
67

5.
47

5.
51

2/
27

2/
32

0/
33

3/
32

1.
83

1.
31

1.
36

1.
24

�0

.0
3

0.
05

0.
24

0.
11

�0
.5
0

�0
.3
1

0.
05

�0
.2
6

PA

N
A
S-

Po
si
ti
ve

.8
6
.9
1
.9
2
.9
1

28
.8
7

30
.9
7

30
.0
2

29
.7
0

8.
41

9.
42

9.
58

9.
43

10
/4
8

10
/5
0

10
/5
0
10
/5
0

1.
56

1.
38

1.
61

1.
40

�0

.0
4

�0
.0
4

�0
.0
6

0.
07

�0
.6
5

�0
.6
4

�0
.6
4

�0
.7
3

PA

N
A
S-
N
eg
at
iv
e

.8
5

.9
3

.9
3
.9
3

19
.6
2

19
.2
8

18

.1
6

19
.8
7

7.
83

9.
34

9.
14

9.
44

10
/4
4

10
/5
0

10
/4
9

10
/4
8
1.
56
1.
41
1.
41
1.
61

0.
89

0.
94

1.
22

0.
85

0.
09

0.
01

0.
85

�0

.2
3

N
ot
e.
Ex
p
er
im
en
ta
l
co
n
d
it
io
n
s
co
m
b
in
ed
w
it
h
in
st
ud
ie
s,

1:
n
¼
22
5;

2:
n
¼
93
2;
3:
n
¼
47
6;
4:
n
¼
89
2.

a:
C
ro
n
b
ac
h
’s

al
p
h
a;

KR
:
Ku

d
er
-R
ic
h
ar
d
so
n
-2
0

A
lp
h
a;

VI
F:

V
ar
ia
n
ce

In
fla
ti
on

Fa
ct
or

(M
ul
ti
co
lli
n
ea
ri
ty
);

D
TA

:
D
ea
th

Th
ou

g
h
t
A
cc
es

si
b
ili
ty

;
SF
FA

:
St
at
e
Se
lf-
Fo
rg
iv
in
g
Fe
el
in
g
s
an
d
A
ct
io
n
s;

SF
B:

Se
lf-
Fo
rg
iv
in
g
Be
lie
fs
;
H
FS
:
H
ea
rt
la
n
d
Fo
rg
iv
en
es
s

Sc
al
e;

PP
RS
:

Pe
rc
ei
ve
d

Pe
rs
on

al
Re
sp
on

si
b
ili
ty
Sc
al
e;

C
D
SI
I:
Re
vi
se
d
C
au
sa
l

D
im
en
si
on

Sc
al
e;

PO
S:

Pe
rc
ei
ve
d

O
ff
en
se

Se
ve
ri
ty
;
SS
G
S:

St
at
e
Sh
am

e
an
d

G
ui
lt

Sc
al
e:

G
ui
lt

Su
b
sc
al
e
(R
:
G
ui
lt

su
b
sc
al
e
re
ve
rs
ed

in
Ex
p
er
im
en
t
1)
;
C
BS
:

C
on

ci
lia
to
ry

Be
h
av
io
rs

Sc
al
e;

PF
-V
ic
ti
m
:
Pe
rc
ei
ve
d

Fo
rg
iv
en
es
s
fr
om

V
ic
ti
m
;
ES
:
Ef
fo
rt

Sc
al
e;
RS
ES
:
Ro
se
n
b
er
g
Se
lf-
Es
te
em

Sc
al
e;
M
-C

SD
S:

M
ar
lo
w
-C
ro
w
n
e
So
ci
al

D
es
ir
ab
ili
ty

Sc
al
e;
PA

N
A
S:
Po
si
ti
ve

an
d
N
eg
at
iv
e
A
ff
ec
t
Sc
h
ed
ul
e.

†S
in
g
le
-i
te
m

sc
al
e.

� E
xp
er
im
en
ts

3
an
d
4
N
eu
tr
al
D
TA

co
n
d
it
io
n
s
ex
cl
ud

ed
in

D
TA
p
ro
p
er
ti
es

p
re
se
n
te
d
.

350 J. M. MCCONNELL

by Hall and Fincham (2008; i.e., The victim has for-
given me and A higher power has forgiven me) rated
on a 5-point Likert-type scale from 1 (not at all) to 5
(completely). These two items are treated as two separ-
ate single-item measures. I used only the perceived
victim forgiveness (PF-Victim) item in the analyses
because Experiment 1 was concerned primarily with
self-forgiveness for interpersonal offenses. PF-Victim
has positively correlated to self-forgiveness, which
supported its convergent validity (Hall & Fincham,
2008). PF-Victim had a mean of 3.41 (SD ¼ 1.27) in
Experiment 1.

Death awareness packet. The following packet was
given to participants ostensibly as data for another
research project in the order presented. These meas-
ures and the experimental manipulation were not
counterbalanced to ensure that each participant had
consistent times between the experimental manipula-
tion and filling in the remaining measures.

Cover story instructions. To conceal the experimental
manipulation, the research assistant told the partici-
pants, “Another researcher has asked us to collect
data for them because they are under a time crunch.
The following is a number of measures the researcher
asked us to have you fill in. … Please fill in this
packet as well.”

Narcissism. The 40-item Narcissistic Personality
Inventory (NPI-40; Raskin & Terry, 1988) was used as
filler for the cover story, and it was not used in the
main statistical analyses. The NPI-40 has demon-
strated adequate levels of internal consistency, 13-
week test–retest reliability, and construct validity as
evidenced by confirmatory factor analysis in previous
samples (del Rosario & White, 2005; Raskin & Terry,
1988). The NPI-40 is a forced-choice questionnaire
whereby the more narcissistic item increases partic-
ipants’ total score (e.g., I am much like everybody else
vs. I am an extraordinary person). The internal con-
sistency was adequate (Kuder-Richardson-20 ¼ .81;
M ¼ 15.98, SD ¼ 6.35) in Experiment 1.

Personality. The 44-item Big Five Inventory (John,
Donahue, & Kentle, 1991) also was used as filler for
the cover story, and it too was not used in the main
statistical analyses. Each of the 44-item Big Five
Inventory’s five domains of personality—Openness to
Experience, Conscientiousness, Extraversion,

Agreeableness, and Neuroticism—has had adequate
internal consistency, test–retest reliabilities, and con-
vergent validity with the NEO Five-Factor Inventory
and the Revised NEO Personality Inventory in previ-
ous samples (John, Naumann, & Soto, 2008;
Rammstedt & John, 2007). Each item is measured on a
5-point Likert-type scale from 1 (disagree strongly) to 5
(agree strongly). The internal consistency for Openness
(a ¼ .77; M¼ 36.29, SD¼ 6.50), Conscientiousness
(a ¼ .75; M¼ 31.19, SD¼ 5.26), Extraversion (a ¼ .88;
M¼ 28.05, SD¼ 6.86), Agreeableness (a ¼ .82; M¼ 34.91,
SD¼ 6.13), and Neuroticism (a ¼ .87; M¼ 23.58,
SD¼ 7.04) were adequate in Experiment 1.

Threat Condition. The threat condition manipulation
consisted of the Mortality Attitudes Personality Survey
(Rosenblatt et al., 1989). This manipulation asks par-
ticipants to respond to two open-ended questions
regarding either death (experimental) or dental pain
(comparison): “Please briefly describe the emotions
that the thought of (your own death | dental pain)
arouses in you” and “Jot down, as specifically as you
can, what you think will happen to you as you physic-
ally (die | experience dental pain) and once you (are
physically dead | have experienced dental pain).” The
Mortality Attitudes Personality Survey is reportedly an
effective manipulation and is the most widely used in
TMT research (Burke et al., 2010; Hayes et al., 2010).
It also yields equivalent effect sizes to other MS
manipulations such as death videos and subliminal
messages (Burke et al., 2010). A neutral condition was
not included in this study because it does not have
larger effect sizes compared to a negative stimulus
such as dental pain (Burke et al., 2010; Steinman &
Updegraff, 2015).

Distractor-delay measures. The following three meas-
ures were given as distractor-delays in counterbal-
anced orders, and I used the first two to rule out
confounding variables.

Social desirability. The Marlowe-Crowne Social
Desirability Scale (M-C SDS; Crowne & Marlowe,
1960) contains 33 true–false items (e.g., I never resent
being asked to return a favor). The M-C SDS
total score increases as participants endorse socially
desirable responses. The M-C SDS has demonstrated
desirable levels of internal consistency reliability and
4-week test–retest reliability in a previous sample
(Crowne & Marlowe, 1960). The M-C SDS has posi-
tively correlated to the Lie validity scale of the

BASIC AND APPLIED SOCIAL PSYCHOLOGY 351

Minnesota Multiphasic Personality Inventory, which
indicated convergent validity (Crowne & Marlowe,
1960). Experiment 1’s sample had a similar mean and
standard deviation (M ¼ 13.71, SD ¼ 5.02) as other
undergraduate samples (see Andrews & Meyer, 2003).
The internal consistency also was adequate
(Kuder–Richardson–20 ¼ .77) in Experiment 1.

Affect. The Positive and Negative Affective Schedule
(PANAS; Watson, Clark, & Tellegen, 1988) has 20
items (e.g., excited, scared) rated on a 5-point Likert-
type scale, from 1 (very slightly or not at all) to 5
(extremely), that assesses positive and negative affect in
the present moment. The Positive and Negative sub-
scales have each demonstrated adequate internal con-
sistency in a previous sample (Watson et al., 1988).
The Positive and Negative subscales have positively
correlated with mood inventories assessing positive or
negative affect, respectively, which supported the con-
vergent validity of the PANAS (Watson et al., 1988).
The internal consistency for Positive Affect (a ¼ .86;
M ¼ 28.87, SD¼ 8.41) and Negative Affect (a ¼ .85;
M ¼ 19.62, SD¼ 7.83) were adequate in Experiment 1.

Puzzle task. Participants were given 3 min to fill in a
word search puzzle used in TMT research to serve as
a distractor-delay (Greenberg, Pyszczynski, Solomon,
Simon, & Breus, 1994).

Mediating measure: DTA. To assess participants’
DTA, I used a word-fragment completion task
(Greenberg et al., 1994). The task consists of 25
words, each missing two letters. Six of these words
can be completed as either death-related (i.e., buried,
dead, grave, killed, skull, coffin) or neutral words (e.g.,
skill). The word-fragment completion task is scored
by tallying up the number of words that participants
complete as death oriented (i.e., 0–6). Putting aside
the complications in assuming nonconscious death
activations even exist, the various convergent findings
in multiple languages support the construct validity of
the word-completion DTA task (Hayes et al., 2010).
DTA had an overall mean of 1.66 (SD ¼ 1.09) in
Experiment 1 with similar means for the MS experi-
mental group (M ¼ 1.65, SD ¼ 1.05) and the DP com-
parison group (M ¼ 1.67, SD ¼ 1.12).

Deception. To reintroduce the remaining criterion
and moderating variables related to self-forgiveness,
the research assistant stated, “Whoops. I messed up
the packets for today. These were supposed to be with

the first study about the event you recalled. Please fill
in these as well.”

Criterion measures. The following two measures
made up the criterion latent-variable “self-forgiveness”
and were given in counterbalanced orders.

Self-forgiveness. The State Self-Forgiveness Scale
(SSFS; Wohl, DeShea, & Wahkinney, 2008) consists of
two subscales—Self-Forgiving Feelings and Actions
(SFFA) and Self-Forgiving Beliefs (SFB)—that have
both demonstrated desirable levels of internal consist-
ency reliabilities in a previous sample (Wohl et al.,
2008). All 17 items are quantified with a 4-point
Likert-type scale, that is, 1 (not at all) to 4 (com-
pletely). The SSFS positively correlated to depression
in a previous study, which supported its convergent
validity (Wohl et al., 2008). The internal consistencies
for SFFA (a ¼ .87; M ¼ 19.06, SD ¼ 5.75) and SFB
(a ¼ .90; M ¼ 26.80, SD ¼ 6.18) were adequate in
Experiment 1.

Guilt. The State Shame and Guilt Scale (SSGS;
Marschall, Sanfter, & Tangney, 1994) consists of 15
items assessing guilt (e.g., I feel remorse, regret),
shame (e.g., I feel worthless, powerless), and pride
(e.g., I feel proud) quantified on a 5-point Likert-type
scale, that is, 1 (not feeling this way at all) to 5 (feeling
this way very strongly). Each subscale of the SSGS has
demonstrated desirable levels of internal consistency
reliabilities in a previous sample (Tangney & Dearing,
2002). The SSGS has good content validity because it
is derived from a large body of theoretical and empir-
ical literature on self-conscious emotions (Tangney &
Dearing, 2002). McConnell et al. (2012) found that
the Guilt subscale was positively correlated to concili-
atory behavior and negatively correlated to self-for-
giveness, which supported its convergent validity. I
excluded administration of the pride subscale in all
experiments. In Experiment 1, I reversed scored the
Guilt scale (GuiltR) to conceptually pair it with state
self-forgiveness. The internal consistency for GuiltR

was adequate (a ¼ .79; M ¼ 11.68, SD ¼ 4.33) in
Experiment 1.

Additional moderating measure: Effort. Although
effort was conceptually a moderator in Experiment 1,
I chose to administer it last because of its relevance to
the process of accomplishing self-forgiveness
(McConnell, 2015). In other words, I gave the meas-
ure last to retrospectively assess the extent that

352 J. M. MCCONNELL

participants felt they had put forth effort toward self-
forgiveness. I used two items modeled after Fisher
and Exline (2006; i.e., How much effort did you take
to reduce your negative feelings? How much time did
you take to reduce your negative feelings?). The Effort
Scale (ES) is quantified on an 11-point Likert-type
scale (0 [no effort] to 10 [great effort]), and the two
items are averaged together to quantify participants’
effort. The ES has demonstrated adequate intercorrela-
tions and internal consistency in a previous sample
(Fisher & Exline, 2006). The ES has positively corre-
lated to repentance, humbling change, and self-for-
giveness, which supported its convergent validity
(Exline et al., 2011; Fisher & Exline, 2006). The
internal consistency was adequate (a ¼ .75; M ¼ 10.43,
SD ¼ 5.09) in Experiment 1.

Demographic questionnaire. I used a demographic
questionnaire asking about participants’ age, gender,
ethnicity, religious affiliation, marital status, level of
education, offense date, and repeat offending.

Results and discussion

Missing data
Excluding cases with missing data is problematic
because it reduces sample size and introduces biases
by modifying the sample means, standard errors, and
standard deviations (Rubin, 1987). There were 51
(.001%) missing datum points within the SSFS, CDSII,
SSGS, M-C SDS, PANAS, and NPI–40 with no more
than two missing datum points per item. Therefore, I
concluded that multiple imputation would introduce
minimal statistical biases. I used AMOS 16.0
(Arbuckle, 2007) to conduct multiple imputation with
10 iterations on each separate subscale to generate val-
ues for the missing data. I then aggregated the 10 iter-
ations of generated values into one data set
for analyses.

Responsibility
Minimizing responsibility (e.g., denying, externalizing)
can masquerade as self-forgiveness on the guilt and
self-forgiveness measures I used in Experiment 1. To
make sure minimizing was a minimal extraneous fac-
tor, I classified the participants’ levels of taking
responsibility with a two-step cluster analysis, one-
way multivariate analysis of variance (MANOVA),
and distribution outliers with SPSS 23.0. Taking
responsibility was operationally defined as having
higher degrees of personal control and lower degrees
of external control. Due to unsatisfactory internal

consistency, I dropped Causality (a ¼ .58) and
Stability (a ¼ .35) from analyses. Using External
Control and Personal Control, two-step cluster ana-
lysis elicited three clusters with good cluster quality as
determined by cohesion and separation (see Table 3).
The lowest responsibility cluster (External: M ¼ 15.54,
SD ¼ 5.97; Personal: M ¼ 12.70, SD ¼ 4.01) had 54
members. The moderate responsibility cluster
(External: M ¼ 14.19, SD ¼ 3.75; Personal: M ¼ 21.80,
SD ¼ 2.53) had 86 members. The highest responsibility
cluster (External: M ¼ 5.79, SD ¼ 2.44; Personal:
M ¼ 24.87, SD ¼ 2.45) had 85 members. A one-way
MANOVA Wilks’s Lambda indicated that clusters had
different levels of external and personal con-
trol (gp

2 ¼ .61).
On external control and personal control, there

were four (1.78%) and 10 (4.44%) participant outliers
beyond 2 SD, respectively. Although there were 54
(24.00%) people included in the lowest responsibility
cluster, only three of these (5.56%) were considered 2
SD outliers on both external and personal control.
Moreover, responsibility was very minimally related to
self-forgiveness (SFFA, External: r ¼ .02; SFB, External:
r ¼ �.08; SFFA, Personal: r ¼ �.03; SFB, Personal:
r ¼ .07) and GuiltR (External: r ¼ �.03; Personal:
r ¼ .02). I concluded that retention of outliers would
not bias the results any more than their exclusion.

Intercorrelation matrices
I created intercorrelation matrices with SPSS 23.0 for
the entire sample (see Table 4), the experimental
group (see Table 5), and the comparison group (see
Table 6).

Main analyses
I used Mplus 6.11 and SPSS 23.0 to test the medi-
ation, moderated-mediation, and interactional-moder-
ated-mediation hypotheses.

Mediation hypothesis. To test the mediation hypoth-
esis of threat condition predicting the latent variable
“self-forgiveness” (i.e., SFFA, SFB, and GuiltR)
through DTA (see Figure 2), I used Mplus 6.11 to
perform a bootstrapped path analysis. The

Table 3. Clusters for entire sample in Experiment 1.
n M SD Range Skewness Kurtosis

Low responsibility cluster: External 54 15.54 5.97 3/27 �0.10 �0.59
Low responsibility cluster: Personal 54 12.70 4.01 3/18 �0.86 0.07
Mod responsibility cluster: External 86 14.19 3.75 8/23 0.35 �0.72
Mod responsibility cluster: Personal 86 21.80 2.53 17/27 0.52 �0.50
High responsibility cluster: External 85 5.79 2.44 3/11 0.61 �0.51
High responsibility cluster: Personal 85 24.87 2.45 17/27 �1.16 0.68
Note. n ¼ 225.

BASIC AND APPLIED SOCIAL PSYCHOLOGY 353

bootstrapped mediation model with 5,000 bootstrap
draws elicited adequate fit, v2(4) ¼ 2.80, v2/df ¼ .70;
comparative fit index =1.00, Tucker–Lewis index
=1.01, root mean square error of approximation ¼ .00;
Akaike information criterion =4601.48, Bayesian infor-
mation criterion =4649.30, but had relations of min-
imal effect sizes (see Table 7): (a) Threat condition
very minimally impacted DTA in the unexpected dir-
ection (b ¼ �.02, b ¼ �.01, R2 = .00), (b) threat condi-
tion minimally increased self-forgiveness (b ¼ 1.66,
b ¼ .15, R2 = .02), and (c) DTA very minimally
impacted self-forgiveness in the unexpected direction
(b ¼ �.04, b ¼ �.01, R2 = .00). The M-C SDS
(gp

2 ¼ � .00) and PANAS–Positive and Negative
(gp

2 ¼ � .00) had near-zero differences between
groups, which indicated neither social desirability nor
affect impacted the results of the mediation model.

Overall, there was a minimal direct effect of MS on
self-forgiveness, but mediation through DTA was
near-zero and in the opposite of the hypothesized dir-
ection. Mediation was not supported.

Moderated-mediation hypothesis. To test the moder-
ated-mediation hypothesis (see Figure 3), I used
Mplus 6.11 to perform a bootstrapped path analysis.
The bootstrapped mediation model with 5,000 boot-
strap draws elicited marginal fit, v2(20) ¼ 43.22, v2/
df ¼ 2.16; comparative fit index ¼ .98, Tucker–Lewis
index ¼ .97, root mean square error of approx-
imation ¼ .07; Akaike information criterion =14335.49,
Bayesian information criterion =14622.44, and had
relations of minimal effect sizes. Again, threat condi-
tion very minimally impacted DTA in the unexpected
direction (b ¼ �.02, b ¼ �.01, R2 = .00) and threat
condition minimally increased self-forgiveness when
the four moderators were considered in the model
(b ¼ 1.54, b ¼ .15, R2 = .02). Individual interactions on
DTA predicting self-forgiveness were small or very
minimal (see Table 7): POS (b ¼ .23, b ¼ .04, R2 =
.00), CBS (b ¼ �.05, b ¼ �.01, R2 = .00), PF-Victim
(b ¼ .88, b ¼ .17, R2 = .03), and ES (b ¼ .10, b ¼ .02,
R2 = .00). Post hoc moderation analyses conducted
with Hayes’s (2013) SPSS 23.0 PROCESS Macro also
revealed interactions that yielded approximately zero
or minimal changes to DTA predicting SFFA, SFB, or
GuiltR individually. In cases of minimal moderation,
the direct effect of DTA on self-forgiveness remained
near-zero and/or in the unexpected direction. There
were no more than 3.89 raw data point changes on
SFFA, SFB, or GuiltR across all values of DTA and
individual moderators (see Table 7). Although

interactions were small or very minimal, they are triv-
ial given the range of scores on SFFA, SFB, and
GuiltR (see Table 2).

Overall, the direct effect of DTA on self-forgive-
ness was minimal and in the opposite of the
hypothesized direction, and individual variables
played little role in moderating this unexpected dir-
ect effect. The results of the moderated-mediation
model further supported that DTA did not meditate
the minimal direct effect of MS on self-forgiveness
despite inclusion of four variables known to
usually effect the self-forgiveness process (see
McConnell, 2015).

Interactional-moderated-mediation hypothesis. To
evidence interactional-moderated-mediation, the experi-
mental group must have more DTA and self-forgiveness
than the comparison group. A one-way analysis of vari-
ance (ANOVA) with threat condition predicting DTA
indicated there was a near-zero difference between MS
(M= 1.65, SD ¼ 1.05) and DP (M = 1.67, SD ¼ 1.12;
gp

2 ¼ � .00). A one-way MANOVA with threat condi-
tion predicting a linear combination of SFFA, SFB, and
GuiltR indicated there was a minimal increase on self-
forgiveness (gp

2 ¼ .02). Post hoc discriminant function
analysis revealed one function (k ¼ .023) that did not
effectively discriminate between groups (56.40% cor-
rectly classified, 53.30% cross-validated). Post hoc one-
way ANOVAs with threat condition predicting each
dependent variable separately again revealed minimal
increases on SFFA (gp

2 ¼ .02), SFB (gp2 ¼ .01), and
GuiltR (gp

2 ¼ .01). These small effects represented no
more than 1.65 raw data point mean differences
between MS (see Table 8; SFFA: M =19.89, SD ¼ 5.30;
SFB: M =27.42, SD ¼ 5.64; GuiltR: M =12.17, SD ¼ 4.33)
and DP (SFFA: M = 18.24, SD ¼ 6.08; SFB: M =26.19,
SD ¼ 6.63; GuiltR: M =11.19, SD ¼ 4.30) on self-forgive-
ness measures. Because there was a near-zero difference
between experimental groups for DTA, I did not test
separate interactional-moderated-mediation models for
each experimental condition. Of note, there were small
or very minimal intercorrelations among MS, DTA, and
self-forgiveness within and between experimental condi-
tions (see Tables 4, 5, and 6).

Overall, there was a near-zero difference between
threat conditions on DTA and MS minimally increased
self-forgiveness. Near-zero effects for DTA are unsup-
portive of interactional-moderated-mediation. These
results replicate the unsupportive findings of the medi-
ation and the moderated-mediation models.

354 J. M. MCCONNELL

Ta
b
le

4.
St
an
d
ar
d
iz
ed

Pe
ar
so
n
an
d
Sp
ea
rm

an
’s
Rh

o
in
te
rc
or
re
la
ti
on

s
in

Ex
p
er
im
en
t
1
fo
r
b
ot
h
ex
p
er
im
en
ta
l
co
n
d
it
io
n
s
co
m
b
in
ed
.

(1
)

(2
)

(3
)

(4
)

(5
)

(6
)

(7
)

(8
)

(9
)

(1
0)

(1
1)

(1
2)

(1
3)

(1
4)

(1
5)

(1
6)

(1
7)

(1
8)

(1
9)

(2
0)

(2
1)

(2
2)

(2
3)

(2
4)

C
on

d
it
io
n
p

(1
)

1
D
TA

(2
)

�.
01

1
SF
FA

(3
)

.1
5

.0
0

1
SF
B

(4
)
.0
8

�.
04

.7
4

1
G
ui
lt
R

(5
)

.1
3

�.
07

.5
9

.4
6

1
PO

S
(6
)

�.
03

�.
02

�.
31

�.
30

�.
38

1
C
BS

(7
)
.0
0
.0
4
�.
30

�.
22

�.
33

.1
3

1
PF
-V
ic
ti
m

(8
)

.1
0

.1
4

.0
0

.0
5

.0
4

�.
36

.2
9

1
ES

(9
)
.1
3
.0
6

�.
16

�.
15

�.
31
.1
3
.3
2
�.
02

1
M
-C

SD
S

(1
0)
.0
5

�.
14

.2
4

.1
8

.1
6

�.
09

.0
4
.1
0
.0
5

1
PA

N
A
S-
Po
si
ti
ve

(1
1)
.0
1
�.
03

.0
9

.0
6

�.
06

�.
10

�.
02
.0
5
.2
5
.2
4
1
PA
N
A
S-
N
eg
at
iv
e
(1
2)
.0
4

�.
05

�.
26

�.
26

�.
27

.2
2

.0
1
�.
22
.1
8

�.
13

.2
6

1
C
D
SI
I:
C
au
sa
lit
y

(1
3)
.0
8
.0
6
�.
10

�.
18

.0
2
�.
01
.0
4
.0
2
�.
07
�.
14
�.
07
.0
6

1
C
D
SI
I:
Ex
te
rn
al

(1
4)
.0
1
�.
07
.0
2

�.
08

�.
03
�.
02
�.
09
�.
15
�.
01
.0
1

�.
00

.0
5

�.
28

1
C
D
SI
I:
Pe
rs
on

al
(1
5)

�.
06
.0
9
�.
03
.0
7
.0
2

�.
12

.1
1

.2
3
�.
01
�.
05
�.
01
�.
10
.2
4

�.
44

1
C
D
SI
I:
St
ab
ili
ty

(1
6)
.0
3
�.
08
�.
05
�.
02
�.
12
.1
0
�.
01

�.
21

.1
1
�.
08
.0
4
.2
0
�.
10
�.
01
�.
07

1
N
PI
-4
0

(1
7)
�.
16
�.
06
�.
01
.0
7
.0
2
�.
01
�.
03
�.
06
.1
4
.0
1
.1
8
.0
2
�.
10
�.
05
.0
5
�.
01

1
BF
I-
44
:
O
p
en
n
es
s

(1
8)
�.
02
�.
01
.0
5
.0
4
.0
7
�.
04
.0
9
�.
02
.0
4
.0
7
.1
1
�.
03

.2
1

�.
10
.1
2
�.
04
.1
6

1
BF
I-
44
:
C
on

sc
ie
n
ti
ou

sn
es
s

(1
9)
.0
0
�.
03
.0
6
.1
1
.0
3
�.
01
.1
3
.1
1
.0
9

.3
7

.2
1
�.
13
�.
12
�.
01
.0
6
.0
8
.1
0
�.
04

1
BF
I-
44
:
Ex
tr
av
er
si
on

(2
0)
.0
2
�.
03
�.
04
.0
8
�.
12
.0
7
.1
5
�.
04
.1
5
.1
3

.3
0

.0
1
�.
07
�.
15
.1
1
.0
5
.4
8
.1
2
.1
6

1
BF
I-
44
:
A
g
re
ea
b
le
n
es
s

(2
1)
�.
02
�.
07
.1
9
.2
1
�.
01
�.
07
.2
2
.2
0
.1
5
.5
3
.2
4

�.
17

�.
13
�.
09
.0
7
�.
03
�.
10
.1
7
.2
3
.2
7

1
BF
I-
44
:
N
eu
ro
ti
ci
sm

(2
2)
.0
5
.0
3
�.
31
�.
22

�.
19

.2
5
.0
2
�.
10
�.
04

�.
48

�.
24

.3
5
.1
6
.0
8
�.
07
.0
0
�.
10
�.
10

�.
29

�.
19

�.
40

1
D
ay
s
Si
n
ce

O
ff
en
se
(2
3)
.1
2
.1
2
.0
0
�.
01
.0
0
�.
07
.0
3
.1
4
.0
1
�.
03
.0
5
.0
0
.0
9
�.
01
�.
06
.0
4
�.
06
�.
06
�.
08
.0
4
.0
9
.0
3

1
Ti
m
es

C
om

m
it
te
d
(2
4)
�.
03
�.
02
.0
1
.0
3
.0
6
�.
05
.0
3
.0
5
.0
6

�.
11

.0
6
�.
08
.1
0
.0
4
.0
4
�.
03
.0
4
.0
8
.0
9
.0
4
.0
0
�.
08
�.
01
1

N
ot
e.
n
¼
22
5.

p
:
Sp
ea
rm

an
’s
Rh

o
C
or
re
la
ti
on

;
D
TA

:
D
ea
th
Th
ou

g
h
t
A
cc
es
si
b
ili
ty
;
SF
FA

:
St
at
e
Se
lf-
Fo
rg
iv
in
g
Fe
el
in
g
s
an
d
A
ct
io
n
s;
SF
B:

Se
lf-
Fo
rg
iv
in
g
Be
lie
fs
;
G
ui
lt
R
:
SS
G
S:
G
ui
lt
it
em

s
re
ve
rs
ed
;
PO

S:
Pe
rc
ei
ve
d
O
ff
en
se

Se
ve
ri
ty
;
C
BS
:
C
on

ci
lia
to
ry
Be
h
av
io
rs
Sc
al
e;
PF
-V
ic
ti
m
:
Pe
rc
ei
ve
d
Fo
rg
iv
en
es
s
fr
om
V
ic
ti
m
;
ES
:
Ef
fo
rt
Sc
al
e;
M
-C
SD
S:
M
ar
lo
w
-C
ro
w
n
e
So
ci
al
D
es
ir
ab
ili
ty
Sc
al
e;

PA
N
A
S:

Po
si
ti
ve
an
d

N
eg
at
iv
e
A
ff
ec
t
Sc
h
ed
ul
e;

C
D
SI
I:
Re
vi
se
d
C
au
sa
l
D
im
en
si
on

Sc
al
e;
N
PI
-4
0:

N
ar
ci
ss
is
ti
c
Pe
rs
on

al
it
y
In
ve
n
to
ry
;
BF
I-
44
:
Bi
g
Fi
ve

In
ve
n
to
ry
.

BASIC AND APPLIED SOCIAL PSYCHOLOGY 355

Ta
b
le

5.
St
an
d
ar
d
iz
ed

Pe
ar
so
n
in
te
rc
or
re
la
ti
on

s
in

Ex
p
er
im
en
t
1
fo
r
m
or
ta
lit
y
sa
lie
n
ce

co
n
d
it
io
n
.

(1
)
(2
)
(3
)
(4
)
(5
)
(6
)
(7
)
(8
)
(9
)
(1
0)
(1
1)
(1
2)
(1
3)
(1
4)
(1
5)
(1
6)
(1
7)
(1
8)
(1
9)
(2
0)
(2
1)
(2
2)
(2
3)
D
TA
(1
)
1
SF
FA
(2
)
.0
1
1
SF
B
(3
)
�.
03
.7
1
1
G
ui
lt
R
(4
)
�.
05

.4
7

.3
8
1
PO

S
(5
)

�.
02

�.
39

�.
28
�.
40
1
C
BS
(6
)
.1
0
�.
18
�.
18

�.
32

.0
6
1
PF
-V
ic
ti
m
(7
)
.2
2
.1
3
.0
5
.1
6

�.
45

.2
6
1
ES
(8
)
�.
03
.0
8
.0
9
�.
24
.0
7
.3
6
�.
04
1
M
-C
SD
S
(9
)
�.
12
.2
9
.1
9
.1
0
�.
22
.0
9
.1
8
.2
8
1
PA
N
A
S-
Po
si
ti
ve
(1
0)
�.
02
.2
7
.1
7
.0
5
�.
22
�.
01
.0
5
.2
7
.2
4
1
PA
N
A
S-
N
eg
at
iv
e
(1
1)
�.
13
�.
26
�.
27

�.
25

.1
2
.0
3
�.
19
.1
2
�.
13

.3
3

1
C
D
SI
I:
C
au
sa
lit
y
(1
2)
.0
9
�.
19
�.
19
�.
10
.0
3
.1
2
.0
0
�.
17
�.
19
.0
2
.1
0
1
C
D
SI
I:
Ex
te
rn
al
(1
3)
.0
0
.0
7
�.
12
�.
07
�.
03
�.
06
�.
05
.0
9
.0
3
.0
5
.0
6

�.
34

1
C
D
SI
I:
Pe
rs
on

al
(1
4)

.0
7
�.
06
.0
6
.0
8
�.
12
.0
9
.1
9
�.
18
�.
10
.0
3
�.
16
.2
7

�.
49

1
C
D
SI
I:
St
ab
ili
ty
(1
5)
�.
07
�.
11
�.
02
�.
09
.1
9
.0
9

�.
20

.1
7
�.
06
�.
02
.0
7
�.
13
�.
05
�.
19
1
N
PI
-4
0
(1
6)
.0
4
.1
4
.2
3
.0
8
.0
2
�.
14
�.
10
.1
5
.0
4
.2
4
.0
1
�.
10
�.
02
.0
0
�.
01
1
BF
I-
44
:
O
p
en
n
es
s
(1
7)
.0
4
.1
7
.0
8
.1
5
�.
14
.0
2
.1
5
�.
09
.1
5
.1
6
�.
09
.2
9

�.
23

.2
7
�.
07
.1
1
1
BF
I-
44
:
C
on
sc
ie
n
ti
ou
sn
es
s
(1
8)
�.
05
.1
2
.1
4
�.
01
.0
2
.1
4
�.
02
.1
9
.3
7
.2
1
�.
18
�.
15
.0
9
�.
09
.0
6
.1
9
�.
01
1
BF
I-
44
:
Ex
tr
av
er
si
on
(1
9)
.0
7
.0
4
.1
9
�.
08
.0
4
.0
7
�.
12
.1
5
.1
1
.2
6
.0
2
�.
01
�.
11
.1
1
.0
8
.4
8
.0
5
.1
2
1
BF
I-
44
:
A
g
re
ea
b
le
n
es
s
(2
0)
.0
0
.2
6
.2
4
�.
05
�.
14
.2
5
.2
5
.3
3

.6
1

.2
6
�.
19
�.
12
�.
09
.0
6
.0
8
�.
06
.2
2
.1
9
.2
8
1
BF
I-
44
:
N
eu
ro
ti
ci
sm
(2
1)
�.
06
�.
45
�.
30
�.
24
.4
1
.0
2
�.
13
�.
13

�.
46

�.
21
.3
5
.2
5
.0
2
�.
01
�.
05
�.
04
�.
23
�.
32
�.
11
�.
38
1
D
ay
s
Si
n
ce
O
ff
en
se
(2
2)
.0
9
�.
04
�.
03
�.
11
.0
4
.0
0
.0
1
.0
4
.0
1
.0
4
.1
0
�.
01
.0
4
�.
16
.0
6
�.
13
�.
12
�.
10
.0
0
.1
2
�.
03
1
Ti
m
es
C
om
m
it
te
d
(2
3)
.2
1
.0
2
�.
06
�.
02
�.
03

��
.1
6

.0
9
�.
08
.0
0
.1
7
.0
4
.1
3
.1
4
.0
5
�.
22
.2
2
.1
8
.0
3
�.
02

��
.1
1

.0
7
.0
3
1
N
ot
e.

n
¼
11
2.

D
TA
:
D
ea
th
Th
ou
g
h
t
A
cc
es
si
b
ili
ty
;
SF
FA
:
St
at
e
Se
lf-
Fo
rg
iv
in
g
Fe
el
in
g
s
an
d
A
ct
io
n
s;
SF
B:
Se
lf-
Fo
rg
iv
in
g
Be
lie
fs
;
G
ui
lt
R
:
SS
G
S:
G
ui
lt
it
em
s
re
ve
rs
ed
;
PO
S:
Pe
rc
ei
ve
d
O
ff
en
se
Se
ve
ri
ty
;
C
BS
:
C
on
ci
lia
to
ry
Be
h
av
io
rs
Sc
al
e;

PF
-V
ic
ti
m
:
Pe
rc
ei
ve
d
Fo
rg
iv
en
es
s
fr
om

V
ic
ti
m
;
ES
:
Ef
fo
rt
Sc
al
e;

M
-C
SD
S:
M
ar
lo
w
-C
ro
w
n
e
So
ci
al
D
es
ir
ab
ili
ty
Sc
al
e;
PA
N
A
S:
Po
si
ti
ve

an
d
N
eg
at
iv
e
A
ff
ec
t
Sc
h
ed
ul
e;

C
D
SI
I:
Re
vi
se
d
C
au
sa
l
D
im
en
si
on
Sc
al
e;
N
PI
-4
0:
N
ar
ci
ss
is
ti
c
Pe
rs
on
al
it
y
In
ve
n
to
ry
;
BF
I-
44
:
Bi
g
Fi
ve
In
ve
n
to
ry
.

356 J. M. MCCONNELL

Ta
b
le

6.
St
an
d
ar
d
iz
ed

Pe
ar
so
n
in
te
rc
or
re
la
ti
on
s
in

Ex
p
er
im
en
t
1
fo
r
d
en
ta
l
p
ai
n
co
n
d
it
io
n
.

(1
)
(2
)
(3
)
(4
)
(5
)
(6
)
(7
)
(8
)
(9
)
(1
0)
(1
1)
(1
2)
(1
3)
(1
4)
(1
5)
(1
6)
(1
7)
(1
8)
(1
9)
(2
0)
(2
1)
(2
2)
(2
3)
D
TA
(1
)
1
SF
FA
(2
)
�.
01
1
SF
B
(3
)
�.
05
.7
5
1
G
ui
lt
R
(4
)
�.
10

.6
9

.5
1

1
PO
S
(5
)
�.
03
�.
24
�.
31

�.
35

1
C
BS
(6
)
�.
01
�.
40
�.
27
�.
35
.1
9
1
PF
-V
ic
ti
m
(7
)
.0
7
�.
13
.0
4
�.
11
�.
27
.3
1
1
ES
(8
)
.1
4
�.
32
�.
32

�.
37

.1
8
.3
1
.0
2
1
M
-C
SD
S
(9
)
�.
17
.1
9
.1
6
.2
0
.0
3
.0
0
.0
1
�.
15
1
PA
N
A
S-
Po
si
ti
ve
(1
0)
�.
03
�.
08
�.
04
�.
18
.0
1
�.
04
.0
4
.2
3
.2
4
1
PA
N
A
S-
N
eg
at
iv
e
(1
1)
.0
3
�.
28
�.
26
�.
30
.3
2
�.
02
�.
25
.2
7
�.
14
.1
9
1
C
D
SI
I:
C
au
sa
lit
y
(1
2)
.0
3
�.
04
�.
19
.1
2
�.
04
�.
03
.0
4
.0
2
�.
11
�.
16
.0
2
1
C
D
SI
I:
Ex
te
rn
al
(1
3)
�.
13
�.
03
�.
05
.0
0
�.
01
�.
12
�.
27
�.
10
�.
02
�.
06
.0
4
�.
22
1
C
D
SI
I:
Pe
rs
on
al
(1
4)
.1
0
�.
01
.0
8
�.
02
�.
12
.1
3
.2
6
.1
3
�.
01
�.
04
�.
04
.2
2
�.
39
1
C
D
SI
I:
St
ab
ili
ty
(1
5)
�.
09
.0
0
�.
02
�.
14
.0
2
�.
10
�.
22
.0
7
�.
11
.1
0
.2
7
�.
07
.0
4
.0
3
1
N
PI
-4
0
(1
6)
�.
15
�.
08
�.
03
.0
0
�.
04
.0
7
.0
0
.0
9
�.
01
.1
3
.0
5
�.
08
�.
08
.0
9
�.
01
1
BF
I-
44
:
O
p
en
n
es
s
(1
7)
�.
06
�.
06
.0
1
.0
1
.0
5
.1
5
�.
18
.1
6
.0
0
.0
5
.0
4
.1
4
.0
4
�.
02
�.
02
.2
1
1
BF
I-
44
:
C
on
sc
ie
n
ti
ou
sn
es
s
(1
8)
�.
01
.0
0
.0
9
.0
7
�.
03
.1
3
.2
3
.0
1
.3
7
.2
2
�.
08
�.
10
�.
10
.1
9
.1
0
.0
3
�.
07
1
BF
I-
44
:
Ex
tr
av
er
si
on
(1
9)
�.
12
�.
11
�.
01
�.
17
.0
9
.2
1
.0
3
.1
7
.1
5
.3
4
.0
0
�.
12
�.
20
.1
1
.0
2
.5
0
.1
9
.1
8
1
BF
I-
44
:
A
g
re
ea
b
le
n
es
s
(2
0)
�.
13
.1
4
.1
9
.0
3
�.
02
.1
9
.1
5
.0
1
.4
6
.1
8
�.
15
�.
14
�.
08
.0
7
�.
13
�.
14
.1
2
.2
6
.2
6
1
BF
I-
44
:
N
eu
ro
ti
ci
sm
(2
1)
.1
1
�.
21
�.
17
�.
17
.1
3
.0
1
�.
09
.0
4

�.
50

�.
26
.3
5
.0
8
.1
3
�.
11
.0
4
�.
14
.0
1
�.
27
�.
26

�.
41

1
D
ay
s
Si
n
ce
O
ff
en
se
(2
2)
.1
5
.0
2
�.
01
.0
8
�.
17
.0
6
.2
6
.0
1
�.
07
.0
6
�.
12
.0
5
�.
06
.0
4
.0
3
.0
3
.0
2
�.
06
.0
6
.0
7
.0
7
1
Ti
m
es
C
om
m
it
te
d
(2
3)
�.
07
.0
2
.0
6
.1
0
�.
07
.0
7
.0
6
.0
9
�.
15
.0
5
�.
12
.1
2
.0
3
.0
4
.0
0
.0
0
.0
8
.1
2
.0
5
.0
2
�.
11
�.
01
1
N
ot
e.

n
¼
11
3.

D
TA
:
D
ea
th
Th
ou
g
h
t
A
cc
es
si
b
ili
ty
;
SF
FA
:
St
at
e
Se
lf-
Fo
rg
iv
in
g
Fe
el
in
g
s
an
d
A
ct
io
n
s;
SF
B:
Se
lf-
Fo
rg
iv
in
g
Be
lie
fs
;
G
ui
lt
R
:
SS
G
S:
G
ui
lt
it
em
s
re
ve
rs
ed
;
PO
S:
Pe
rc
ei
ve
d
O
ff
en
se
Se
ve
ri
ty
;
C
BS
:
C
on
ci
lia
to
ry
Be
h
av
io
rs
Sc
al
e;
PF
-V
ic
ti
m
:
Pe
rc
ei
ve
d
Fo
rg
iv
en
es
s
fr
om
V
ic
ti
m
;
ES
:
Ef
fo
rt
Sc
al
e;
M
-C
SD
S:
M
ar
lo
w
-C
ro
w
n
e
So
ci
al
D
es
ir
ab
ili
ty
Sc
al
e;
PA
N
A
S:
Po
si
ti
ve
an
d
N
eg
at
iv
e
A
ff
ec
t
Sc
h
ed
ul
e;
C
D
SI
I:
Re
vi
se
d
C
au
sa
l
D
im
en
si
on
Sc
al
e;
N
PI
-4
0:
N
ar
ci
ss
is
ti
c
Pe
rs
on
al
it
y
In
ve
n
to
ry
;
BF
I-
44
:
Bi
g
Fi
ve
In
ve
n
to
ry
.

BASIC AND APPLIED SOCIAL PSYCHOLOGY 357

Summary
The results of Experiment 1 were unsupportive of the
mediation, moderated-mediation, and interactional-
moderated-mediation hypotheses. First, MS did not
lead to greater DTA. In fact, DTA was on average
slightly larger in the DP comparison group. Second,
more DTA negligibly led to less self-forgiveness, and
moderations by perceived offense severity, conciliatory
behaviors, perceived forgiveness, and effort were near-
zero or inconsequently small. Third, although there
was a minimal direct effect of MS on self-forgiveness,
the effect accounted for only 2% of the variance in
self-forgiveness and approximately 1 raw data point
increases. This small effect is negligibly supportive of
existential motivations for self-forgiveness in context
of TMT’s all-encompassing claims.

The unsupportive and negligible findings of
Experiment 1 could be explained by experimental or
measurement artifacts. First, because the offense recall
threat did not experimentally manipulate perceived
offense severity and guilt intensity or measure effect on

self-esteem, participants’ anxiety-buffers may have still
remained strong prior to the MS manipulation. Second,
several measurement artifacts of the experimental
sequence may have impacted the results. The use of
conciliatory behaviors, perceived forgiveness, and nar-
cissism measures may have unintentionally strength-
ened the participants’ anxiety buffer. Also, the
measurement-of-mediation design may have inadvert-
ently exposed the comparison group to death-related
stimuli while filling in the DTA measure. Third, using
a convenience sample of young adults who self-selected
to speak about wrongdoings may have biased the
results. Finally, because I did not use a model-building
approach to investigate an experimental-causal-chain,
the findings of Experiment 1 cannot directly evaluate
mediation due to its measurement-of-mediation design.

For these reasons, I dismantled the experimental
sequence into three follow-up experimental-causal-
chain studies that addressed the aforementioned con-
founding factors and more fully isolated TMT Hy1,
TMT Hy2, and TMT Hy3 where feasible. When

Figure 2. Mediation path analysis model. Note. Standardized beta weights presented. SFFA: Self-Forgiving Feelings and Actions;
SFB: Self-Forgiving Beliefs; GuiltR: State Shame and Guilt Scale: Guilt (reversed).

Table 7. Summary table of mediation and moderated-mediation model results for Experiment 1.
Mediation Moderated-Mediation

Predictor Moderator Criterion b b R2 b b R2 DR2 Draw

MS j DP on DTA �.02 �.01 .00 �.02 �.01 .00 – –
MS j DP on SF 1.66 .15 .02 1.54 .15 .02 – –
DTA on SF �.04 �.01 .00 �.04 �.01 .00 – –
DTA with POS on SF – – – .23 .04 – .00 –

CBS on SF – – – �.05 �.01 – .00 –
PF-Victim on SF – – – .88 .17 – .03 –
ES on SF – – – .10 .02 – .00 –

DTA with POS on SFFA – – – – – – .00 –
SFB – – – – – – .00 –
GuiltR – – – – – – .00 –

CBS on SFFA – – – – – – .00 –
SFB – – – – – – .01 3.89
GuiltR – – – – – – .00 –

PF-Victim on SFFA – – – – – – .02 1.96
SFB – – – – – – .01 2.30
GuiltR – – – – – – .00 –

ES on SFFA – – – – – – .00 –
SFB – – – – – – .02 2.15
GuiltR – – – – – – .01 3.57

Note. n ¼ 225. b: unstandardized beta; b: standardized beta; Draw: max raw data change in interaction on criterion variable; MS: Mortality Salience;
DP: Dental Pain Salience; SFFA: State Self-Forgiving Feelings and Actions; SFB: State Self-Forgiving Beliefs; GuiltR: SSGS:Guilt items reversed; SF: “Self-for-
giveness” latent variable (SFFA, SFB, and GuiltR); POS: Perceived Offense Severity; CBS: Conciliatory Behaviors Scale; PF-Victim: Perceived Forgiveness
from Victim; ES: Effort Scale.

358 J. M. MCCONNELL

relevant, subsequent studies included two levels of the
anxiety-buffer threat, a self-esteem measure, and a
neutral word-completion measure. In addition, I
excluded the conciliatory behaviors, perceived forgive-
ness, and narcissism measures and replaced the CDSII
with a more parsimonious measure. Moreover, I con-
ducted the follow-up studies via online modules to
use a more generalizable population and remove any
self-selection bias. Of note, participants were randomly
assigned to one of three experiments without repetitive
participation, and a postexperiment questionnaire
affirmed they had never before participated in a self-
forgiveness or death-related study. Finally, responsibil-
ity, perceived offense severity, and guilt were concep-
tualized both as part of the anxiety-buffer
manipulation itself and as manipulation checks rather
than moderating or criterion variables.

Experiment 2

To address alternative explanations of Experiment 1, I
conducted a study designed to test TMT Hy1 while
controlling for TMT Hy3. Specifically, I provided two
levels of the anxiety-buffer threat condition (mild vs.
bad) along with manipulation checks that were trailed
by an MS manipulation. I utilized a mild threat com-
parison group because a no-offense threat control
group would be conceptually unrelated to state self-
forgiveness. I used a randomized, posttest-only, dou-
ble-blind four-group factorial design (mild MS vs.
mild DP vs. bad MS vs. bad DP). I expected that MS
would predict greater levels of self-forgiveness than

DP salience (TMT Hy1) and that this effect would
occur with the bad offense recall to a larger degree
than the mild offense recall (TMT Hy3).

Method

Participants
A convenience sample of 932 U.S. citizens who were
at least 18 years old participated online via Qualtrics
in exchange for approximately $1.5 USD compensa-
tion. See Table 1 for a summary of demographic
characteristics.

Procedure
Participants first completed an informed consent and
then were randomly assigned to one of four condi-
tions (mild MS vs. mild DP vs. bad MS vs. bad DP;
n ¼ 233 each). Participants then recalled and briefly
wrote about a mild or bad interpersonal wrongdoing
committed by them within the past year. As part of
the anxiety-buffer manipulation, participants filled in
counterbalanced measures of responsibility, perceived
offense severity, and guilt, and then finished with a
self-esteem manipulation check. Participants then
completed the death awareness packet with distractor-
delays all identical to Experiment 1 but without the
NPI-40 and the overlooked-measures deception. Last,
participants filled in a state self-forgiveness scale and
subsequently completed a demographic questionnaire.
See Figure 4 for a visual depiction of the experimen-
tal sequence.

Figure 3. Moderated-mediation path analysis model. Note. Standardized beta weights presented. SFFA: Self-Forgiving Feelings and
Actions; SFB: Self-Forgiving Beliefs; GuiltR: State Shame and Guilt Scale: Guilt (reversed).

BASIC AND APPLIED SOCIAL PSYCHOLOGY 359

Materials
I excluded the CDSII, CBS, PF-Victim, NPI-40, DTA,
ES, and deception from Experiment 2 (cf. Figures 1
and 4) but retained the distractor-delay measures,
SSFS, and demographic questionnaire. The following
highlight other major modifications or additions. See
Table 2 for Experiment 2 scale properties.

Offense packet. The offense packet was modified by
including a two-level offense recall, a different meas-
ure of responsibility along with the POS and SSGS in
counterbalanced orders, and a self-esteem manipula-
tion check.

Offense recall. Participants were instructed,

Please recall a recent event in the past year in which
you did something (mildly wrong | wrong) to someone.
Recall something you did that was (fairly minor |
pretty bad). Only recall something that most people
would believe is really [not that bad (e.g., borrowing a
pen and not returning it) | a pretty bad thing to do
(e.g., betraying someone, neglecting something very
important, physical altercation) and NOT something
that is [a pretty bad thing to do (e.g., betraying
someone, neglecting something very important,
physical altercation) | fairly minor (e.g., borrowing a
pen and not returning it)]. This should be something
that you take responsibility for doing. For a moment,
visualize in your mind the events and the interactions
you may have had in this situation. Try to visualize the
person and recall what happened. Now please briefly
describe the event in three to five sentences.

My research assistant and I reviewed offense recalls
to ensure participants adequately responded to the
prompt with a mild or bad interpersonal offense.

Responsibility. The CDSII was replaced with the more
parsimonious Perceived Personal Responsibility Scale
(PPRS; Fisher & Exline, 2006). The PPRS is a five-item
measure on a 10-point Likert-type scale from 1 (com-
pletely disagree) to 10 (completely agree) that assesses the
extent to which people believe they are responsible for
an offense (e.g., “This was clearly my fault”). The scale
has demonstrated adequate levels of internal consistency
reliability in a previous sample (Fisher & Exline, 2006).

Perceived offense severity. The offense packet retained
the POS.

Guilt. The offense packet now included the SSGS Guilt
and Shame subscales in Experiment 2. The Guilt sub-
scale was used for analyses.

Self-esteem. The Rosenberg Self-Esteem Scale (RSES;
Rosenberg, 1965) was added as a manipulation check
at the end of the offense packet. The RSES is a 10-item
measure on a 4-point Likert-type scale from 0 (strongly
disagree) to 3 (strongly agree) that quantifies self-esteem
(e.g., “I feel that I have a number of good qualities”).
The RSES is a widely used measure of self-esteem, and

Table 8. Summary table of experimental results.
Experiment 1 Experiment 2

Mild Bad

MS DP MS DP MS DP

M SD M SD g2p k % M SD M SD M SD M SD g
2
p

DTA 1.65 1.05 1.67 1.12 .00 – –
SF – – – – .02 .02 56.40 SFFA 21.51 5.82 22.53 5.24 20.71 5.59 20.67 6.02 .02
SFFA 19.89 5.30 18.24 6.08 .02 – – Offense Recall .01
SFB 27.42 5.64 26.19 6.63 .01 – – Death Threat .00
GuiltR 12.17 4.33 11.19 4.30 .01 – – Interaction .00

SFB 28.09 6.46 29.14 5.58 27.46 6.57 26.42 7.26 .02
Offense Recall .02
Death Threat .00
Interaction .00

Experiment 3 (DTA conditions only) Experiment 4 (DTA conditions only)

Predictor Criterion Mild DTA Bad DTA Mild Neutral Bad Neutral

M SD M SD b b� R2 M SD M SD M SD M SD g2p
RSES on DTA 19.92 6.53 1.88 .99 �.02 �.14 .02 DTA 1.63 .99 1.94 1.04 – – – – .02
DTA on HFS 1.88 .99 29.00 7.28 �1.20 �.16 .03 SFFA 23.09 5.18 20.75 5.98 – – – – .04

SFB 29.57 5.71 26.95 6.87 – – – – .04
Predictor Criterion

M SD M SD b b� R2
DTA on SFFA 1.79 1.02 21.92 5.71 �.70 �.13 .02
DTA on SFB 1.79 1.02 28.26 6.44 �.95 �.15 .02

Note. gp
2: partial eta-squared; k: discriminant function eigenvalue; %: percentage classified in post hoc discriminant analysis; b: unstandardized beta;

b: standardized beta; SFFA: State Self-Forgiving Feelings and Actions; SFB: State Self-Forgiving Beliefs; GuiltR: SSGS:Guilt items reversed; SF: all linear
combination variables in multivariate analysis of variance (SFFA, SFB, and GuiltR); RSES: Rosenberg Self-Esteem Scale; DTA: Death Thought Accessibility.

360 J. M. MCCONNELL

the scale has demonstrated adequate levels of internal
consistency reliability and construct validity in previous
samples (Robins, Hendin, & Trzesniewski, 2001).

Death awareness packet
The death awareness packet was identical to
Experiment 1 except for the exclusion of the NPI-40.

Results and discussion

Manipulation check
I used SPSS 23.0 for all analyses in Experiment 2.
There were no missing data. A factorial ANOVA indi-
cated the PPRS was different among mild MS
(M ¼ 35.75, SD¼ 11.53), mild DP (M ¼ 35.73,
SD¼ 10.92), bad MS (M ¼ 37.04, SD¼ 11.80), and bad
DP (M ¼ 38.15, SD¼ 10.62; gp2 ¼ .01) with a small
main effect for offense recall type (gp

2 ¼ .01) and a
near-zero effect for death-threat condition
(gp

2 ¼ � .00). A factorial ANOVA revealed that the
POS was different among mild MS (M ¼ 14.58,
SD¼ 2.92), mild DP (M ¼ 14.43, SD¼ 3.00), bad MS
(M ¼ 15.93, SD¼ 3.61), and bad DP (M ¼ 16.09,
SD¼ 3.54; gp2 ¼ .05) with a medium main effect for
offense recall type (gp

2 ¼ .05) and a near-zero effect for
death-threat condition (gp

2 ¼ � .00). A factorial
ANOVA indicated that the SSGS Guilt subscale was
different among mild MS (M ¼ 15.03, SD¼ 5.78), mild
DP (M ¼ 14.52, SD¼ 5.76), bad MS (M ¼ 17.61,
SD¼ 5.57), and bad DP (M ¼ 17.79, SD¼ 5.52;
gp

2 ¼ .06) with a medium main effect for offense recall
type (gp

2 ¼ .06) and a near-zero effect for death-threat
condition (gp

2 ¼ � .00). A factorial ANOVA specified

the RSES was different among mild MS (M ¼ 18.73,
SD¼ 6.56), mild DP (M ¼ 19.70, SD¼ 6.08), bad MS
(M ¼ 18.12, SD¼ 6.25), and bad DP (M ¼ 17.40,
SD¼ 6.01; gp2 ¼ .02) with a small main effect for
offense recall type (gp

2 ¼ .01) and a near-zero effect for
death-threat condition (gp

2 ¼ � .00). As expected, the
offense recall manipulation increased levels of responsi-
bility, perceived offense severity, and guilt, as well as
decreased self-esteem for the bad offense recall group.

Main analyses
A factorial ANOVA revealed the SFFA was different
among mild MS (M ¼ 21.51, SD ¼ 5.82), mild DP
(M ¼ 22.53, SD ¼ 5.24), bad MS (M ¼ 20.71,
SD ¼ 5.59), and bad DP (M ¼ 20.67, SD ¼ 6.02;
gp

2 ¼ .02); unexpectedly, there was a near-zero effect
for death-threat condition (gp

2 ¼ � .00), but there
was a small main effect for offense recall type
(gp

2 ¼ .01). There was a near-zero interaction effect
(gp

2 ¼ � .00). A factorial ANOVA found that the
SFB was different among mild MS (M ¼ 28.09,
SD ¼ 6.46), mild DP (M ¼ 29.14, SD ¼ 5.58), bad MS
(M ¼ 27.46, SD ¼ 6.57), and bad DP (M ¼ 26.42,
SD ¼ 7.26; gp2 ¼ .02) with the same unexpected find-
ings for death-threat (gp

2 ¼ �.00) and offense recall
type conditions (gp

2 ¼ .02). There was a near-zero
interaction effect (gp

2 ¼ � .00). The M-C SDS had a
near-zero difference among groups (gp

2 ¼ � .00).
Although PANAS-Positive had a near-zero differ-
ence among groups (gp

2 ¼ � .00), PANAS-Negative
had a small difference among mild MS (M ¼ 18.60,
SD ¼ 8.95), mild DP (M ¼ 18.36, SD ¼ 8.94), bad MS
(M ¼ 19.61, SD ¼ 9.31), and bad DP (M ¼ 20.58,

Additional MaterialsOffense Packet
Death Awareness

Packet
Distractor-Delay

Measures
Dependent Variable

B
A
D

M
I
L
D

M
I
L
D
B
A
D
M
O
R
T
A
L
I
T
Y
S
A
L
I
E
N
C
E
Counterbalanced
Social
Desirability
(M-C SDS)
Mood
(PANAS)
Puzzle Task
Self-Forgiveness
(SSFS)
Demographic
Questionnaire
D
E
N
T
A
L
P
A
I
N
Counterbalanced
Social
Desirability
(M-C SDS)
Mood
(PANAS)
Puzzle Task
Cover
Story
Personality
(BFI-44)
Self-Forgiveness
(SSFS)
Counterbalanced

Responsibility
(PPRS)

Perceived Offense
Severity
(POS)
Guilt
(SSGS)

Self-Esteem

(RSES)

Demographic
Questionnaire
Threat
Condition
(MAPS)

Mortality
Salience

OR
Dental
Pain
Salience

Threat
Condition

Mild
Offense
Recall

OR

Bad
Offense
Recall

Figure 4. Graphical representation of the experimental procedure for Experiment 2.

BASIC AND APPLIED SOCIAL PSYCHOLOGY 361

SD ¼ 10.03, gp2 ¼ .01). Understandably, the small
main effect on PANAS-Negative resulted from
offense recall type (gp

2 ¼ .01) and not from death
threat (gp

2 ¼ � .00). Overall, threat condition had a
near-zero effect on self-forgiveness regardless of
offense recall conditions.

Exploratory analyses
I simultaneously entered PPRS (b ¼ �.04, b ¼ �.08),
POS (b ¼ �.08, b ¼ �.05), SSGS Guilt (b ¼ �.20,
b ¼ �.20), and RSES (b ¼ .48, b ¼ .52) into a multiple
linear regression model predicting SFFA (R2 = .43). I
also simultaneously entered PPRS (b ¼ �.04,
b ¼ �.07), POS (b ¼ �.08, b ¼ �.04), SSGS Guilt
(b ¼ �.15, b ¼ �.14), and RSES (b ¼ .57, b ¼ .54) into
a multiple linear regression model predicting SFB
(R2 ¼ .39). Although PPRS and POS were minimally
related, SSGS guilt was negatively associated with self-
forgiveness to a small to moderate degree and RSES
was positively related to a large degree.

Summary
The results of Experiment 2 suggested that the min-
imal direct effect observed in Experiment 1 was likely
spurious. In Experiment 2, MS had a near-zero effect
on self-forgiveness even in conditions with higher
responsibility taking, more severe offenses, greater
guilt, and lower self-esteem. These results did not sup-
port TMT Hy1 when controlling for TMT Hy3.
Although there were small to moderate differences
between offense recall groups on manipulation checks,
MS should have had an effect on self-forgiveness at
least commensurate with self-esteem levels to support
TMT Hy1. In addition, findings were actually the
exact opposite of TMT Hy3 as self-forgiveness was
higher in the low anxiety-buffer threat condition
rather than the high anxiety-buffer threat condition.
Exploratory analyses revealed that guilt was negatively
associated with self-forgiveness and that self-esteem
was positively related.

Experiment 3

The measurement-of-mediation design of Experiment
1 cannot infer causation. Experimental-causal-chains
more rationally test the temporal sequence implied in
mediation (Kline, 2015; Spencer et al., 2005; Tate,
2015; Trafimow, 2015). Because there is no known
way to directly manipulate DTA without MS or
anxiety-buffer threats, baseline DTA studies are a
second-best alternative (Hayes et al., 2010). To more
rationally investigate the mediational aspect of TMT

Hy2, I conducted a dispositional self-esteem, baseline
DTA, and trait self-forgiveness study. To account for
unintended death-related cognition, I also included a
neutral word-fragment completion task condition. I
used a randomized, double-blind comparison group
design (DTA vs. neutral). I predicted that baseline
DTA would positively relate to trait self-forgiveness
(TMT Hy2) and that dispositional self-esteem would
negatively relate to baseline DTA (TMT Hy3) for the
participants in the DTA word-fragment comple-
tion condition.

Method

Participants
A convenience sample of 476 U.S. citizens who were
at least 18 years old participated online via Qualtrics
in exchange for approximately $1.5 USD remuner-
ation. See Table 1 for a summary of demographic
characteristics.

Procedure
Participants first completed an informed consent and
then were randomly assigned to one of two conditions
(DTA vs. neutral; n ¼ 238 each). Participants then
completed a self-esteem measure, followed by a DTA
or neutral word-fragment completion task and then a
trait self-forgiveness measure. Subsequently, partici-
pants filled in social desirability and affect measures
in counterbalanced orders. Finally, they completed a
demographic questionnaire. See Figure 5 for a visual
depiction of the experimental sequence.

Materials
I used the RSES, DTA word-completion task, M-C
SDS, PANAS, demographic questionnaire, and the fol-
lowing measures in Experiment 3. See Table 2 for
Experiment 3 scale properties.

Neutral word-fragment completion task. To account
for unintended death-related cognition, I used a neu-
tral word-fragment completion task that consisted of
25 words. Each word was missing two letters and
could not be completed in death-related ways.

Trait Self-Forgiveness. The Heartland Forgiveness
Scale (HFS; Thompson et al., 2005) is an 18-item
measure on a 7-point Likert-type scale from 1 (almost
always false of me) to 7 (almost always true of me)
that assesses dispositional forgiveness of self, others,
and situations. I used the six-item HFS self-forgive-
ness subscale (e.g., “With time I am understanding of

362 J. M. MCCONNELL

myself for mistakes I’ve made”) to measure trait self-
forgiveness. The HFS has demonstrated desirable lev-
els of internal consistency reliability and convergent
validity in a previous sample (Thompson et al., 2005).

Results and discussion

Main analyses
I used SPSS 23.0 for all analyses in Experiment 3. There
were no missing data. To explore the relations among
dispositional self-esteem, baseline DTA, and trait self-
forgiveness, I analyzed data only from the DTA word-
fragment completion condition, as the neutral condition
did not have a chance to fill in death-related words.
There was a small negative relation with RSES predicting
baseline DTA in a linear regression model (b¼ �.02,
b ¼ �.14, R2 = .02). Unexpectedly, there was a small
negative relation rather than a positive relation when
baseline DTA predicted HFS in a linear regression
model (b ¼ �1.20, b ¼ �.16, R2 = .03). A one-way
ANOVA indicated there was a near-zero difference in
HFS across word-fragment completion task conditions
(gp

2 ¼ � .00), which indicated that filling in the DTA
measure in and of itself minimally impacted trait self-
forgiveness measurement. In addition, M-C SDS
(gp

2 ¼ � .00), PANAS-Positive (gp2 ¼ � .00), and
PANAS-Negative (gp

2 ¼ � .00) had near-zero differences
across word-fragment completion task conditions.
Overall, the results are mixed. The RSES had a small
negative relation with baseline DTA, but unexpectedly

baseline DTA was negatively related to HFS to a
small degree.

Exploratory analysis
RSES positively related to HFS in linear regression
model to a large degree (b ¼ .73, b ¼ .69, R2 ¼ .47).

Summary
The results of Experiment 3 are consistent with the
unsupportive findings of Experiment 1 and 2.
Although Experiment 3 found a small negative rela-
tion between dispositional self-esteem and baseline
DTA, self-esteem only accounted for 2% of the vari-
ance in DTA. This finding is negligibly supportive of
TMT Hy3 in consideration that TMT theorists believe
the anxiety buffer is the main regulator of DTA.
Alternatively, dispositional self-esteem may not have
been low enough to impact baseline DTA more. Yet
the small relation between self-esteem and DTA did
not correspond with the predicted relation between
baseline DTA and trait self-forgiveness. Experiment 3
actually found the opposite of TMT Hy2. Higher levels
of baseline DTA were associated with less trait self-
forgiveness. Exploratory analyses revealed self-esteem
was positively related with self-forgiveness.

Experiment 4

To experimentally test TMT Hy3 independent of TMT
Hy1, I provided an identical anxiety-buffer threat as in

Additional MaterialsCriterion VariableBaseline Measure Additional MaterialsBaseline Measure

D
T
A

Counterbalanced
Social
Desirability
(M-C SDS)
Mood
(PANAS)
Death-Thought
Accessibility
(DTA Word
Completion
Task)
Demographic
Questionnaire

N
E
U
T
R
A
L

Counterbalanced
Social
Desirability
(M-C SDS)
Mood
(PANAS)

Neutral Word
Completion Task

Self-Esteem
(RSES)
Demographic
Questionnaire

Trait
Self-Forgiveness

(HFS)

Trait
Self-Forgiveness
(HFS)

Figure 5. Graphical representation of the experimental procedure for Experiment 3.

BASIC AND APPLIED SOCIAL PSYCHOLOGY 363

Experiment 2, excluded the MS manipulation, and
measured DTA after a delay. Some TMT Hy1 research
has supported that DTA effects deteriorate over time
when threats are not explicitly death related
(Steinman & Updegraff, 2015); however, few to no
studies have systematically examined distractor-delay
effects in the context of an anxiety-buffer threat
(Webber, Zhang, Schimel, & Blatter, 2016). Therefore,
I chose to keep a distractor-delay to keep methodo-
logical consistency between Experiments 1 and 4, but
I measured the time delay in seconds to account for
its effect. To speculate about DTA mediation, I also
included a self-forgiveness measure and neutral word-
fragment completion task conditions to account for
unintended death-related cognition. I used a random-
ized, posttest-only, double-blind four-group factorial
design (mild DTA vs. mild neutral vs. bad DTA vs.
bad neutral) combined with a measurement-of-medi-
ation procedure. I expected that the bad offense recall
would predict greater DTA than the mild offense
recall (TMT Hy3). I also predicted that DTA would be
positively related to self-forgiveness in the DTA con-
ditions (TMT Hy2) and self-forgiveness levels would
be equivalent across the DTA and neutral conditions.

Method

Participants
A convenience sample of 892 U.S. citizens who were
at least 18 years old participated online via Qualtrics
in exchange for approximately $1.5 USD compensa-
tion. See Table 1 for a summary of demographic
characteristics.

Procedure and materials
Participants first completed an informed consent and
then were randomly assigned to one of four condi-
tions (mild DTA vs. mild neutral vs. bad DTA vs. bad
neutral; n ¼ 223 each). They then completed the exact
same offense packet as Experiment 2 followed by the
same distracter-delay measures in counterbalanced
orders. Next, they filled in either the DTA or the neu-
tral word-fragment completion task, trailed by the
SSFS, and finally a demographic questionnaire. See
Figure 6 for a visual depiction of the experimental
sequence, and see Table 2 for Experiment 4
scale properties.

Results and discussion

Manipulation check
I used SPSS 23.0 for all analyses in Experiment 4.
There were no missing data. A factorial ANOVA indi-
cated that the PPRS was different among mild DTA
(M ¼ 34.85, SD ¼ 11.09), mild neutral (M ¼ 36.26,
SD ¼ 11.77), bad DTA (M ¼ 37.81, SD ¼ 11.10), and
bad neutral (M ¼ 37.22, SD ¼ 10.82, gp2 ¼ .01) with a
small main effect for offense recall type (gp

2 ¼ .01)
and a near-zero effect for word-fragment completion
condition (gp

2 ¼ � .00). A factorial ANOVA revealed
that the POS was different among mild DTA
(M ¼ 13.73, SD ¼ 3.35), mild neutral (M ¼ 14.25,
SD ¼ 3.30), bad DTA (M ¼ 15.87, SD ¼ 3.64), and bad
neutral (M ¼ 16.43, SD ¼ 3.47; gp2 ¼ .10) with a
medium main effect for offense recall type (gp

2 ¼ .09)
and a near-zero effect for word-fragment completion
condition (gp

2 ¼ � .00). A factorial ANOVA indicated

Additional MaterialsDependent VariableOffense Packet
Distractor-Delay

Measures
Dependent Variable
B
A
D
M
I
L
D
M
I
L
D
B
A
D
D
T
A
Counterbalanced
Social
Desirability
(M-C SDS)
Mood
(PANAS)
Puzzle Task
Death-Thought
Accessibility
(DTA Word
Completion
Task)
Demographic
Questionnaire
N
E
U
T
R
A
L
Counterbalanced
Social
Desirability
(M-C SDS)
Mood
(PANAS)
Puzzle Task
Neutral Word
Completion Task
Counterbalanced
Responsibility
(PPRS)
Perceived Offense
Severity
(POS)
Guilt
(SSGS)
Self-Esteem
(RSES)
Demographic
Questionnaire
Threat
Condition
Mild
Offense
Recall
OR
Bad
Offense
Recall
Self-Forgiveness
(SSFS)
Self-Forgiveness
(SSFS)

Figure 6. Graphical representation of the experimental procedure for Experiment 4.

364 J. M. MCCONNELL

that the SSGS Guilt subscale was different among
mild DTA (M ¼ 14.31, SD ¼ 5.67), mild neutral
(M ¼ 14.32, SD ¼ 5.84), bad DTA (M ¼ 17.61,
SD ¼ 5.59), and bad neutral (M ¼ 17.86, SD ¼ 5.49,
gp

2 ¼ .08) with a medium main effect for offense recall
type (gp

2 ¼ .08) and a near-zero effect for word-frag-
ment completion condition (gp

2 ¼ � .00). A factorial
ANOVA specified that the RSES was different among
mild DTA (M ¼ 19.51, SD ¼ 6.17), mild neutral
(M ¼ 19.67, SD ¼ 6.62), bad DTA (M ¼ 16.85,
SD ¼ 6.84), and bad neutral (M ¼ 17.10, SD ¼ 6.04,
gp

2 ¼ .04) with a small to medium main effect for
offense recall type (gp

2 ¼ .04) and a near-zero effect
for word-fragment completion condition (gp

2 ¼ � .00).
As with Experiment 2, the offense recall manipulation
had the intended effect on responsibility, perceived
offense severity, guilt, and self-esteem for the bad
offense recall group.

Main analyses
As with Experiment 3, I analyzed only data from the
DTA word-fragment completion condition for
the main analyses. A one-way ANOVA revealed that
the offense recall had a minimal effect on DTA
for the mild DTA (M ¼ 1.63, SD ¼ .99) and bad
DTA conditions (M ¼ 1.94, SD ¼ 1.04, gp2 ¼ .02).
There was a near-zero relation between distractor-
delay seconds (M ¼ 477.44, SD ¼ 321.62, min ¼ 208.58,
max ¼ 3903.15) and DTA in a linear regression model
(b ¼ .00, b ¼ .03, R2 = .00). The M-C SDS had a near-
zero difference among groups (gp

2 ¼ � .00). Although
PANAS-Positive had a near-zero difference among
groups (gp

2 ¼ � .00), PANAS-Negative had a small
difference among mild DTA (M ¼ 18.53, SD ¼ 8.54),
mild neutral (M ¼ 18.19, SD ¼ 9.28), bad DTA
(M ¼ 21.06, SD ¼ 9.94), and bad neutral (M ¼ 21.71,
SD ¼ 9.52, gp2 ¼ .03). Understandably, the small main
effect on PANAS-Negative resulted from offense recall
type (gp

2 ¼ .03) and not from word-fragment comple-
tion condition (gp

2 ¼ � .00). Overall, the offense recall
had a minimal effect on DTA.

Exploratory analyses
To explore if DTA measurement had an effect on
self-forgiveness measurement, I collapsed offense
recall conditions within word-fragment completion
task conditions and then conducted a one-way
ANOVA that indicated SFFA had a near-zero differ-
ence between word-fragment completion task condi-
tions (gp

2 ¼ �.00). Also, a one-way ANOVA indicated
that SFB had a near-zero difference between word-
fragment completion task conditions (gp

2 ¼ �.00).

These findings indicated that filling in the DTA meas-
ure in and of itself minimally impacted self-forgive-
ness measurement.

Within the DTA word-fragment completion task
conditions, a one-way ANOVA indicated SFFA was
greater in the mild offense recall condition
(M ¼ 23.09, SD ¼ 5.18) than the bad offense recall
condition (M ¼ 20.75, SD ¼ 5.98, gp2 ¼ .04). Also, a
one-way ANOVA indicated SFB was greater in the
mild offense recall condition (M ¼ 29.57, SD ¼ 5.71)
than the bad offense recall condition (M ¼ 26.95,
SD ¼ 6.87, gp2 ¼ .04). Unexpectedly, DTA was nega-
tively related to both SFFA (b ¼ �.70, b ¼ �.13, R2 =
.02) and SFB (b ¼ �.95, b ¼ �.15, R2 = .02) to a small
degree in separate linear regression models. Within
the DTA word-fragment completion task conditions, I
simultaneously entered PPRS (b ¼ �.04, b ¼ �.07),
POS (b ¼ �.06, b ¼ �.04), SSGS Guilt (b ¼ �.16,
b ¼ �.16), and RSES (b ¼ .47, b ¼ .55) into a multiple
linear regression model predicting SFFA (R2 = .42). I
also simultaneously entered PPRS (b ¼ .00, b ¼ .00),
POS (b ¼ .00, b ¼ .00), SSGS Guilt (b ¼ �.18,
b ¼ �.16), and RSES (b ¼ .60, b ¼ .62) into a multiple
linear regression model predicting SFB (R2 = .47).
Although PPRS and POS were minimally related or
very minimally related, SSGS Guilt was negatively
associated with self-forgiveness to a small degree and
RSES was positively related to a large degree. Overall,
unexpectedly, DTA was negatively related to self-for-
giveness to a small degree.

Summary
The results of Experiment 4 are consistent with the
unsupportive findings of Experiments 1, 2, and 3.
Although Experiment 4 found a small relation between
offense recall type and DTA, offense recall only
accounted for 2% of the variance in DTA. This corre-
sponded with a mean difference of 0.31 death-related
word completions between the mild and bad offense
recall conditions. Again, this is negligibly supportive
for TMT Hy3 given the central nature of self-esteem as
a buffer against DTA in TMT. Alternatively, the offense
recalls perhaps did not reduce the anxiety buffer
enough for DTA to become preconscious to a higher
degree. Nonetheless, the small relation between offense
recall and DTA did not correspond with the predicted
relation between DTA and self-forgiveness in explora-
tory components of Experiment 4. Measurement of
DTA had a near-zero impact on self-forgiveness meas-
urement, but even if it did inadvertently change self-
forgiveness, higher levels of DTA were associated with
less self-forgiveness. In addition, a more severe anxiety-

BASIC AND APPLIED SOCIAL PSYCHOLOGY 365

buffer threat actually predicted less rather than more
self-forgiveness. These findings are the opposite of
TMT Hy2 and TMT Hy3. Exploratory analyses revealed
that guilt was negatively associated with self-forgiveness
and that self-esteem was positively related.

General discussion

Review of findings

A large body of literature has indicated that self-for-
giveness is related to overall well-being, including
existential well-being (Davis et al., 2015; McConnell,
2015). As proposed by McConnell’s (2015) C-T-E
framework, TMT could potentially explain the basis
for self-forgiveness. With theory and research in both
TMT and self-forgiveness as a rationale, I positioned
self-forgiveness within the mortality-salience (Hy1),
death-thought accessibility (Hy2), and anxiety-buffer
hypotheses of TMT (Hy3).

In Experiment 1, I hypothesized an interactional-
moderated-mediation model among MS, DTA, and
self-forgiveness would be supported. Specifically, I
hypothesized that after an anxiety-buffer threat to
induce guilt for a recent interpersonal offense,
experimentally manipulated MS would increase
self-forgiveness through DTA after a delay, and the
relation between DTA and self-forgiveness would be
moderated by offense severity, conciliatory behav-
iors, perceived forgiveness, and effort to reduce
emotions. Experiment 1 did not support TMT Hy1,
TMT Hy2, or TMT Hy3. Although the mediation,
moderated-mediation, and interactional-moderated-
mediation models were not supported, there was a
minimal direct effect of MS on self-forgiveness that
was negligibly supportive of existential motivations
for self-forgiveness. Because the results of
Experiment 1 may have been disturbed by experi-
mental or measurement artifacts, I followed up
with a three-study experimental-causal-chain. In
Experiment 2, an expanded four-group factorial
experimental manipulation without measurement of
DTA mediation did not support TMT Hy1 while
controlling for TMT Hy3. The MS manipulation had
a near-zero effect on self-forgiveness at both levels
of the anxiety-buffer threat. This suggested that the
small direct effect found in Experiment 1 was spuri-
ous. In Experiment 3, TMT Hy3 was negligibly sup-
ported when higher dispositional self-esteem
predicted lower baseline DTA to a minimal degree,
and the opposite of TMT Hy2 was found when
higher baseline DTA predicted lower trait self-for-
giveness. In Experiment 4, a two-level anxiety buffer

threat with an exploratory measurement of DTA
mediation procedure negligibly supported TMT Hy3

and found the inverse of TMT Hy2. More severe
offense recalls led to minimal increases in DTA and
DTA was associated with less self-forgiveness.

Eternal sunshine of the spotless mind or spotty
empirical findings?

Most perplexing is that in two experiments MS min-
imally increased or very minimally impacted partic-
ipants’ levels of a well-reasoned “distal defense” to
self-esteem (i.e., self-forgiveness) after a delay.
Although MS effects in Experiment 1 may have been
obfuscated by unintended MS exposure in the com-
parison group while measuring DTA (Hayes et al.,
2010; Hayes & Schimel, 2018), the negligible findings
actually were further reduced in Experiment 2 with a
two-level anxiety-buffer threat and no measurement
of DTA. The current studies did not support TMT
Hy1. Also puzzling is that initial MS did not increase
participants’ levels of DTA after a delay as indicated
by TMT and much of its published research (Hayes
et al., 2010; Pyszczynski et al., 1999; Steinman &
Updegraff, 2015). Although baseline and experimen-
tally induced DTA was minimally associated with dis-
positional self-esteem and self-esteem threats in two
experiments, DTA was never associated with more
self-forgiveness. In fact, higher DTA was associated
with less self-forgiveness in three studies. The current
studies along with others directly contradicted TMT
Hy2 (e.g., Hart, 2014; Hughes, 2013; Trafimow &
Hughes, 2012). Likewise confusing is that one-level
and two-level anxiety-buffer threats in multiple
experiments did not result in MS, DTA, or distal
defense effects consistent with TMT and its research
(e.g., Hayes et al., 2010; Pyszczynski et al., 1999,
2015). In experiments with two-level offense recalls,
there was actually higher self-forgiveness with less
severe anxiety-buffer threats. The current studies
failed to support TMT’s anxiety-buffer hypothesis.
Overall, the negligible and unexpected findings of the
current studies have challenged the central assump-
tions of TMT, its most widely used methodologies,
and McConnell’s (2015) C-T-E model of self-forgive-
ness. Existential motivations for self-forgiveness were
unsupported, whereas emotional, social-cognitive, and
offense-related determinants were associated with self-
forgiveness in exploratory zero-order correlation and
multiple linear regression analyses. Specifically,
exploratory correlation and regression analyses
revealed attributions, perceived offense severity, guilt,

366 J. M. MCCONNELL

perceived forgiveness, effort to reduce emotions, neur-
oticism, and self-esteem all had associations with self-
forgiveness consistent with prior research
(McConnell, 2015).

These unsupportive findings have importance for
future experimental existential psychology studies as
well as theoretical refinement of McConnell’s (2015)
C-T-E model of self-forgiveness. Scholars in psych-
ology have a history of ignoring or trivializing negli-
gible findings and confirmatory approaches to
publication. A biased publication approach is not sci-
entific. A truly scientific approach gives equal weight
to negligible findings, disconfimatory evidence, and
alternative explanations (Popper, 1963; Tracey &
Glidden-Tracey, 1999; Trafimow, 2014). The paradox
of scientific progress is that important theoretical
developments follow the dismantling of commonly
accepted and seemingly valid ideas (Kuhn, 1962).
These scientific revolutions begin from falsification
(Kuhn, 1962; Popper, 1963) and from negligible find-
ings (Tracey & Glidden-Tracey, 1999; Trafimow,
2014). Reviewing the history of physics is helpful in
illustrating the importance of valuing negligible find-
ings in psychology and in science more general
(Trafimow, 2014). Toward this aim, Trafimow (2014)
highlighted a famous failed experiment by Michelson
and Morley (1887). Michelson and Morley were
unable to support the previously popular idea that
luminiferous ether carried light waves. They failed to
detect any luminiferous ether, and this gave way for
physicists to accept Einstein’s (1905) groundbreaking
theory of special relativity. In addition, observing
black swans is an apt analogy for scientific falsifica-
tion. How do you know all swans are not white, or
death awareness is not all-encompassing? You falsify
the theory by finding one or more black swans, or
one or more negligible TMT findings. There are black
swans, and there are negligible TMT findings from
experiments with internal validity. The results of the
current studies found no compelling evidence to sup-
port existential motivations for self-forgiveness or
decisive support for general anxiety-buffer or DTA
effects. Rather, the current studies along with prior
experiments (e.g., Hart, 2014; Hughes, 2013;
Trafimow & Hughes, 2012) indicated that TMT’s
body of evidence is inconsistent or even contradictory
to the theory. Either the theory is not truthful or the
theory is not as robust as its proponents
have suggested.

Given the results of the current studies, scholars
can reconcile the extant TMT and self-forgiveness
literature in one of three ways: (a) TMT is not true,

(b) TMT does not reliably elicit DTA and distal
defenses experimentally, or (c) self-forgiveness is not
a distal defense despite TMT declaring death
explains virtually all human cognition, emotion, and
behavior. First, despite its impressive empirical sup-
port, TMT has considerable theoretical and empir-
ical flaws that need to be reconciled (see Martin &
van den Bos, 2014; see also Psychological Inquiry
special issues: Pyszczynski et al., 1997 [Vol. 8, Issue
1] and Martin & Erber, 2006 [Vol. 17, Issue 4]). For
instance, Martin and van den Bos (2014) clarified
that TMT uses inconsistent evidence in confirmatory
ways and ignores alternative explanations. Scholars
will have to complete further conceptual, theoretical,
and empirical work to understand more fully the
veracity of TMT. Second, effect sizes summarized in
meta-analytic reviews have indicated that TMT does
not reliably produce supportive evidence. Although
Burke and colleagues’ (2010) and Steinman and
Updegraff’s (2015) reports provided meta-analytic
support for the MS and DTA hypotheses, researchers
must consider common misconceptions of meta-
analytic reviews (Aguinis, Pierce, Bosco, Dalton, &
Dalton, 2011). For instance, single-effect size sum-
maries cannot provide a full picture of TMT’s publi-
cation biases (cf. Yen & Cheng, 2013), its file drawer
problems (cf. Hart, 2014; Hughes, 2013), and its
effect size ranges (cf. Trafimow & Hughes, 2012).
Indeed, TMT meta-analyses include studies with
near-zero, minimal, and contradictory effect sizes
(cf. Burke et al., 2010; Hart, 2014; Hughes, 2013;
Proulx & Heine, 2008; Steinman & Updegraff, 2015;
Trafimow & Hughes, 2012; Yen & Cheng, 2010,
2013). As TMT’s research findings continue to accu-
mulate in, I hope, an unbiased and scientific fashion,
scholars may be able to come to a deeper under-
standing of the robustness of TMT. Finally, if TMT
is indeed true, then its assumptions seem to suggest
that self-forgiveness should be a distal defense for
self-esteem reductions secondary to guilt. Although
this first attempt to empirically investigate this line
of reasoning was unsupportive and contradicted
TMT in multiple studies, the findings of the current
studies are somewhat inconclusive about TMT’s pos-
sible role in self-forgiveness because MS and anx-
iety-buffer threats did not have robust effects on
self-esteem and DTA, the fundamental mechanisms
believed to bring about distal defenses. On one
hand, perhaps MS and anxiety-buffer threats did not
produce the expected DTA and self-forgiveness
effects because TMT is not true or because TMT
effects are more circumscribed and less powerful

BASIC AND APPLIED SOCIAL PSYCHOLOGY 367

than proposed by its proponents. On the other hand,
perhaps terror management motivations for self-for-
giveness are nonexistent or they were simply not
detected in the current studies because manipula-
tions were not robust enough.

As self-forgiveness research using TMT accumu-
lates, scholars may come to a more definitive
conclusion about existential motivations for self-
forgiveness. However, perhaps a number of alterative
theories to TMT may better explain the origin of
self-forgiveness. Three potential theoretical frame-
works are (a) the stress and coping model (Lazarus
& Folkman, 1984), (b) the integrative theory of psy-
chological defense (Hart, 2014), and (c) the tripartite
conceptual view of meaning in life (George & Park,
2016). The stress and coping model proposed people
use problem-focused and emotion-focused coping to
bring about resolutions to stressful events (Lazarus
& Folkman, 1984). The integrative theory of psycho-
logical defense claims threats to insecurity (death
related or otherwise) activates the security system
comprising self-esteem, attachment, and worldviews
(Hart, 2014). The tripartite conceptual view of
meaning in life integrates various meaning frame-
works, such as the meaning maintenance model
(Heine et al., 2006), with conceptualizations of
meaning in life (i.e., comprehension, purpose, and
mattering; George & Park, 2016). Alternatively, cur-
rent or future attempts at a unifying theory in
psychology may be fruitful (see Trafimow, 2012b;
see also Gaj, 2016). Future research studies can clar-
ify what theoretical frameworks have explanatory
power in why people experience self-forgiveness, and
thereby continue to clarify the C-T-E model of self-
forgiveness proposed by McConnell (2015).

Strengths

There are several methodological strengths worthy of
mention. First, I used all three of TMT’s main
hypotheses in the design of multiple studies. Although
no one study can fully rule out extraneous factors,
an omnibus initial study with a follow-up experimen-
tal-causal-chain dismantling procedure minimizes
alternative explanations. Second, the randomized,
posttest-only, double-blind comparison group and fac-
torial designs alongside counterbalancing of measures
protects well against internal threats to validity
(Shadish et al., 2002). The current series of studies
rules out threats to internal validity such as history,
maturation, testing, selection, mortality, experimenter
bias, and hypotheses guessing. Finally, I utilized three

distractor-delays following the experimental manipula-
tions. Given that similar distractor-delays are what
have supposedly caused the highest effect sizes in
DTA and distal defenses (Burke et al., 2010; Steinman
& Updegraff, 2015), the current studies’ results cannot
be explained by insufficient distractor-delays.

There also are a number of statistical conclusion
validity strengths. First, the breadth of variables
accounted for in Experiment 1 was an important
strength in attempting to account for unexplained
variance and interactions. By using important varia-
bles in the extant self-forgiveness literature, the study
attempted to clarify how interpersonal, offense-spe-
cific, and behavioral variables modify potential exist-
ential motivations for self-forgiveness. Second, a
number of confounds were ruled out, such as con-
struct invalidity (e.g., minimizing offenses), socioemo-
tional variables (e.g., social desirability & affect),
nonrecent offenses, and ceiling or flooring effects.
Third, counterbalancing measures improves the math-
ematical rationale for using mediation approaches that
have experimental manipulations, measurement-of-
mediation, and conceptually based temporal sequences
(Kline, 2015; Preacher et al., 2007; Spencer et al.,
2005; Tate, 2015). Finally, utilization of a large sample
size increases the statistical conclusion validity related
to negligible findings.

Limitations

Although the current series of studies protects well
against traditional threats to statistical and internal
validity, they had some limitations. First, low psycho-
metric internal consistency reliability of perceived
offense severity partially reduced the statistical conclu-
sion validity of the moderation results in Experiment
1 (Kline, 2015). Thus, the results concerning offense
severity should remain cautionary.

Second, there are a number of psychometric and
methodological problems with measuring DTA.
Although word-fragment completion tasks are
widely used in TMT research, questions can be
raised about the extent they actually measure non-
conscious death thoughts, or if nonconscious death
thoughts ever could be measured by anything.
Operationalization of nonobservable constructs must
include sound specification of auxiliary assumptions
that rationally bridge constructs with observable
measures (Trafimow, 2012a). Perhaps DTA meas-
ures, such as word-fragment completion tasks, do
not meet this auxiliary specification requirement. If
this is the case, then the current studies may have

368 J. M. MCCONNELL

actually missed measuring the DTA that TMT
claims. In addition, measuring DTA in Experiment 1
may have contaminated the comparison group with
death-related cognition if participants responded in
death-related ways on the word-fragment completion
task (Hayes et al., 2010; Hayes & Schimel, 2018).
However, death-thought contamination was ruled out
as an extraneous factor when I included neutral word-
fragment completion task conditions in Experiment 4.
In addition, because the measurement-of-mediation
component in Experiments 1 and 4 did not manipulate
DTA, these studies do not directly allow for causal
inferences between DTA and self-forgiveness (Kline,
2015; Spencer et al., 2005; Tate, 2015; Trafimow, 2015).
Measurement-of-mediation designs can only suggest
possible causal mediation at best. Regardless, the inabil-
ity to make causal inferences does not change the cur-
rent studies’ conclusions. After all, findings of
Experiments 1 and 4 did not suggest possible or actual
mechanisms of mediation but actually supported a lack
of mediation. Measurement-of-mediation results also
were reinforced through the unsupportive findings of
the experimental-causal-chain series. If a mediation
between MS and self-forgiveness through DTA existed,
then the results from the hypothetical mediation mod-
els and the experimental-causal-chain would have at
least suggested the hypothesized relations.

Third, Experiment 4 had a distractor-delay that
may have obscured anxiety-buffer threat effects on
DTA. Although there was not a relation between dis-
tractor-delay time and DTA for the time frames
observed, there was no one who had a zero distractor-
delay time. On average, participants experienced a
477.44-s delay across three distracting measures with a
minimum of 208.58-s within the sample. If anxiety-
buffer effects on DTA completely disappear with any
time delay, Experiment 4 would have not detected an
effect. Future TMT research should systematically
investigate various distractor-delay times with offense-
related anxiety-buffer threats.

Fourth, an interaction of selection and treatment in
Experiment 1 limits generalizability because an under-
graduate sample consisting mostly of young, unmarried
White Christians is different than the general popula-
tion in meaningful ways, such as how relevant death is
developmentally. However, three follow-up experiments
using a more generalizable population had similar find-
ings. Nonetheless, future ecologically valid research may
explore TMT and self-forgiveness in populations such
as young people in hospice care and death-row inmates
to further clarify generalizability. After all, the terror of

death may be most relevant for young individuals
whose lives may be cut short (Pyszczynski et al., 2015).

Conclusion

Ethical debates concerning whether self-forgiveness is
warranted are controversial, but correlational research
has suggested that people do seem to benefit from
self-forgiveness in various biopsychosocial domains of
well-being (Davis et al., 2015; McConnell, 2015). Yet
scholars have not fully clarified the nature of why
people experience self-forgiveness. McConnell (2015)
proposed that terror management theory may be a
cogent explanation for the motivation behind self-for-
giveness. However, the results of the current studies
do not support the assumption that MS leads to
increased self-forgiveness, either directly or indirectly
through DTA and various moderators, in the context
of a self-esteem threat of recalling recent interpersonal
offenses. This suggested self-forgiveness is not an
existentially motivated experience. Alternatively, these
studies and other studies (e.g., Hart, 2014; Hughes,
2013; Proulx & Heine, 2008; Trafimow & Hughes,
2012; Yen & Cheng, 2013) challenge the veracity of
TMT itself, and other existentially oriented theories
may explain the origin of self-forgiveness. Future
studies, including replications and extensions, as well
as alternative theoretical frameworks may help eluci-
date why people strive for self-forgiveness and provide
further refinement to McConnell’s (2015) C-T-E
model of self-forgiveness.

Acknowledgments

Portions of this article are from the author’s doctoral disser-
tation completed at Ball State University. Dr. McConnell
expresses his gratitude to Paul M. Spengler, PhD, Stefan�ıa
Ægisd�ottir, PhD, W. Holmes Finch, PhD, Thomas M.
Holtgraves, PhD, David N. Dixon, PhD, Felicia A. Dixon,
PhD, Kristopher J. Preacher, PhD, and David Trafimow,
PhD for their helpful consultation. He also expresses grati-
tude to his research assistants, Tully Roll and
Donna Azcuna.

ORCID

John M. McConnell http://orcid.org/0000-0002-
4742-9854

References

Aguinis, H., Pierce, C. A., Bosco, F. A., Dalton, D. R., &
Dalton, C. M. (2011). Debunking myths and urban

BASIC AND APPLIED SOCIAL PSYCHOLOGY 369

legends about meta-analysis. Organizational Research
Methods, 14, 306–331. doi:10.1177/1094428110375720

Andrews, P., & Meyer, R. G. (2003). Marlowe-Crowne
social desirability scale and short form C: Forensic
norms. Journal of Clinical Psychology, 59, 483–492. doi:
10.1002/jclp.10136

Anglin, S. M. (2014). From avoidance to approach: The
effects of mortality salience and attachment on the motiv-
ation to repair troubled relationships. Personality and
Individual Differences, 66, 86–91. doi:10.1016/
j.paid.2014.03.012

Arbuckle, J. L. (2007). Amos 16.0 user’s guide. Chicago, IL:
Amos Development Corporation.

Arndt, J., Cook, A., Goldenberg, J. L., & Cox, C. R. (2007).
Cancer and the threat of death: The cognitive dynamics
of death thought suppression and its impact on behav-
ioral health intentions. Journal of Personality and Social
Psychology, 92, 12–29. doi:10.1037/0022-3514.92.1.12

Arndt, J., Greenberg, J., Solomon, S., Pyszczynski, T., &
Schimel, J. (1999). Creativity and terror management:
Evidence that creative activity increases guilt and social
projection following mortality salience. Journal of
Personality and Social Psychology, 77, 19–32. doi:10.1037/
0022-3514.77.1.19

Baron, R. M., & Kenny, D. A. (1986). The moderator–me-
diator variable distinction in social psychological
research: Conceptual, strategic, and statistical considera-
tions. Journal of Personality and Social Psychology, 51,
1173–1182. doi:10.1037/0022-3514.51.6.1173

Burke, B. L., Kosloff, S., & Landau, M. J. (2013). Death goes
to the polls: A meta-analysis of mortality salience effects
no political attitudes. Political Psychology, 34, 183–200.
doi:10.1111/pops.12005

Burke, B. L., Martens, A., & Faucher, E. H. (2010). Two
decades of terror management theory: A meta-analysis of
mortality salience research. Personality and Social
Psychology Review, 14, 155–195. doi:10.1177/
1088868309352321

Cox, C. R., Eaton, S., Ekas, N. V., & Van Enkevort, E. A.
(2015). Death concerns and psychological well-being in
mothers of children with autism spectrum disorder.
Research in Developmental Disabilities, 45-46, 229–238.
doi:10.1016/j.ridd.2015.07.029

Cox, C. R., Reid-Arndt, S. A., Arndt, J., & Moser, R. P.
(2012). Considering the unspoken: The role of death cog-
nition in quality of life among women with and without
breast cancer. Journal of Psychosocial Oncology, 30,
128–139. doi:doi:10.1080/07347332.2011.633980.

Crowne, D. P., & Marlowe, D. (1960). A new scale of social
desirability independent of psychopathology. Journal of
Consulting and Clinical Psychology, 24, 349–354. doi:
10.1037/h0047358

Darrell, A., & Pyszczynski, T. (2016). Terror management
theory: Exploring the role of death in life. In L. A.
Harvell & G. S. Nisbett (Eds.), Denying death: An inter-
disciplinary approach to terror management theory
(pp. 1–15). New York, NY: Routledge.

Das, E., Bushman, B. J., Bezemer, M. D., Kerkhof, P., &
Vermeulen, I. E. (2009). How terrorism news reports
increase prejudice against outgroups: A terror manage-
ment account. Journal of Experimental Social Psychology,
45, 453–459. doi:10.1016/j.jesp.2008.12.001

Davis, D. E., Ho, M. Y., Griffin, B. J., Bell, C., Hook, J. N.,
Van Tongeren, D. R., … Westbrook, C. J. (2015).
Forgiving the self and physical and mental health corre-
lates: A meta-analytic review. Journal of Counseling
Psychology, 62, 329–335. doi:10.1037/cou0000063

Davis, D. E., Worthington, E. L., Jr., Hook, J. N., & Hill, P.
C. (2013). Research on religion/spirituality and forgive-
ness: A meta-analytic review. Psychology of Religion and
Spirituality, 5, 233–241. doi:10.1037/a0033637

del Rosario, P. M., & White, R. M. (2005). The Narcissistic
Personality Inventory: Test-retest stability and internal
consistency. Personality and Individual Differences, 39,
1075–1081. doi:10.1016/j.paid.2005.08.001

Echebarria-Echabe, A. (2013). Mortality salience and uncer-
tainty: Similar effects but different processes?. European
Journal of Social Psychology, 43, 185–191. doi:10.1002/
ejsp.1938

Einstein, A. (1905). Zur Elektrodynamik bewegter K€orper.
Annalen Der Physik, 17, 890–921. doi:10.1002/
andp.19053221004

Exline, J. J., DeShea, L., & Holeman, V. T. (2007). Is
apology worth the risk? Predictors, outcomes, and ways
to avoid regret. Journal of Social and Clinical Psychology,
26, 479–504. doi:10.1521/jscp.2007.26.4.479

Exline, J. J., Root, B. L., Suhrida, Y., Martin, A. M., &
Fisher, M. L. (2011). Reparative behaviors and self-for-
giveness: Effects of a laboratory-based exercise. Self and
Identity, 10, 101–126. doi:10.1080/15298861003669565

Fawcett, J. (1993). Analysis and evaluation of nursing theo-
ries. Philadelphia: F. A. Davis.

Fergus, T. A., & Valentiner, D. P. (2012). Terror manage-
ment theory and scrupulosity: An experimental investiga-
tion. Journal of Obsessive-Compulsive and Related
Disorders, 1, 104–111. doi:10.1016/j.jocrd.2012.01.003

Ferrell, B., Otis-Green, S., Baird, P., & Garcia, A. (2014).
Nurses’ responses to requests for forgiveness and the end
of life. Journal of Pain and Symptom Management, 47,
631–641. doi:10.1016/j.jpainsymman.2013.05.009

Fisher, M. L., & Exline, J. J. (2006). Self-forgiveness versus
excusing: The roles of remorse, effort, and acceptance of
responsibility. Self and Identity, 5, 127–146. doi:10.1080/
15298860600586123

Fisher, M. L., & Exline, J. J. (2010). Moving toward self-for-
giveness: Removing barriers related to shame, guilt, and
regret. Social and Personality Psychology Compass, 4,
548–558. doi:10.1111/j.1751-9004.2010.00276.x

Fransen, M. L., Fennis, B. M., Pruyn, A. T. H., & Das, E.
(2008). Rest in peace? Brand-induced mortality salience
and consumer behavior. Journal of Business Research, 61,
1053–1061. doi:10.1016/j.jbusres.2007.09.020

Gaj, N. (2016). Unity and fragmentation in psychology: The
philosophical and methodological roots of the discipline.
New York, NY: Routledge

George, L. S., & Park, C. L. (2016). Meaning in life as com-
prehension, purpose, and mattering: Toward integration
and new research. Review of General Psychology, 20,
205–220. doi:10.1037/gpr0000077

Goldenberg, J. L., Heflick, N. A., & Cooper, D. P. (2008).
The thrust of the problem: Bodily inhibitions and guilt as
a function of mortality salience and neuroticism. Journal
of Personality, 76, 1055–1080. doi:10.1111/j.1467-
6494.2008.00513.x

370 J. M. MCCONNELL

https://doi.org/10.1177/1094428110375720

https://doi.org/10.1002/jclp.10136

https://doi.org/10.1016/j.paid.2014.03.012

https://doi.org/10.1016/j.paid.2014.03.012

https://doi.org/10.1037/0022-3514.92.1.12

https://doi.org/10.1037/0022-3514.77.1.19

https://doi.org/10.1037/0022-3514.77.1.19

https://doi.org/10.1037/0022-3514.51.6.1173

https://doi.org/10.1111/pops.12005

https://doi.org/10.1177/1088868309352321

https://doi.org/10.1177/1088868309352321

https://doi.org/10.1016/j.ridd.2015.07.029

https://doi.org/10.1080/07347332.2011.633980.

https://doi.org/10.1037/h0047358

https://doi.org/10.1016/j.jesp.2008.12.001

https://doi.org/10.1037/cou0000063

https://doi.org/10.1037/a0033637

https://doi.org/10.1016/j.paid.2005.08.001

https://doi.org/10.1002/ejsp.1938

https://doi.org/10.1002/ejsp.1938

https://doi.org/10.1002/andp.19053221004

https://doi.org/10.1002/andp.19053221004

https://doi.org/10.1521/jscp.2007.26.4.479

https://doi.org/10.1080/15298861003669565

https://doi.org/10.1016/j.jocrd.2012.01.003

https://doi.org/10.1016/j.jpainsymman.2013.05.009

https://doi.org/10.1080/15298860600586123

https://doi.org/10.1080/15298860600586123

https://doi.org/10.1111/j.1751-9004.2010.00276.x

https://doi.org/10.1016/j.jbusres.2007.09.020

https://doi.org/10.1037/gpr0000077

https://doi.org/10.1111/j.1467-6494.2008.00513.x

https://doi.org/10.1111/j.1467-6494.2008.00513.x

Greenberg, J., Pyszczynski, T., & Solomon, S. (1986). The
causes and consequences of a need for self-esteem: A ter-
ror management theory. In R. F. Baumeister (Ed.), Public
self and private self (pp. 189–212). New York: Springer-
Verlag.

Greenberg, J., Pyszczynski, T., Solomon, S., Simon, L., &
Breus, M. (1994). The role of consciousness and accessi-
bility of death related thoughts in mortality salience
effects. Journal of Personality and Social Psychology, 67,
627–637. doi:10.1037/0022-3514.67.4.627

Greenberg, J., Solomon, S., Pyszczynski, T., Rosenblatt, A.,
Burling, J., Lyon, D., … Pinel, E. (1992). Why do people
need self-esteem? Converging evidence that self-esteem
serves as an anxiety-buffering function. Journal of
Personality and Social Psychology, 63, 913–922. doi:
10.1037/0022-3514.63.6.913

Hall, J. H., & Fincham, F. D. (2005). Self-forgiveness: The
stepchild of forgiveness research. Journal of Social and
Clinical Psychology, 24, 621–637. doi:10.1521/
jscp.2005.24.5.621

Hall, J. H., & Fincham, F. D. (2008). The temporal course
of self-forgiveness. Journal of Social and Clinical
Psychology, 27, 174–202. doi:10.1521/jscp.2008.27.2.174

Harrison, P. R., & Mallet, R. K. (2013). Mortality salience
motivates the defense of environmental values and
increases collective ecoguilt. Ecopsychology, 5, 36–43. doi:
10.1089/eco.2012.0070

Hart, J. (2014). Toward an integrated theory of psycho-
logical defense. Perspectives on Psychological Science, 9,
19–39. doi:10.1177/1745691613506018

Hayes, A. F. (2013). Introduction to mediation, moderation,
and conditional process analysis: A regression-based
approach. New York, NY: The Guilford Press.

Hayes, J., & Schimel, J. (2018). Unintended effects of meas-
uring implicit processes: The case of death-thought acces-
sibility in mortality salience studies. Journal of
Experimental Social Psychology, 74, 257–269. doi:10.1016/
j.jesp.2017.10.005

Hayes, J., Schimel, J., Arndt, J., & Faucher, E. H. (2010). A
theoretical and empirical review of death-thought accessi-
bility concept in terror management research.
Psychological Bulletin, 136, 699–739. doi:10.1037/
a0020524

Hayes, J., Schimel, J., Williams, T. J., Howard, A. L.,
Webber, D., & Faucher, E. H. (2015). Worldview accom-
modation: Selectively modifying committed beliefs pro-
vides defense against worldview threat. Self and Identity,
14, 521–548. doi:10.1080/15298868.2015.1036919

Heflick, N. A. (2005). Sentenced to die: Last statements and
dying on death row. Omega – Journal of Death and
Dying, 51, 323–336. doi:10.2190/96X8-FLUT-TCLH-EL71

Heine, S. J., Proulx, T., & Vohs, K. D. (2006). The meaning
maintenance model: On the coherence of social motiva-
tions. Personality and Social Psychology Review, 10,
88–110. doi:10.1207/s15327957pspr1002_1

Henriques, G. (2011). A new unified theory of psychology.
New York, NY: Springer

Hirschberger, G., Florian, V., & Mikulincer, M. (2002). The
anxiety buffering function of close relationships:
Mortality salience effects on the readiness to compromise
mate selection standards. European Journal of Social
Psychology, 32, 609–625. doi:10.1002/ejsp.110

Hughes, J. S. (2013). [Tests of TMT death thought suppres-
sion and rebound hypothesis]. Unpublished raw data.

Iverach, L., Menzies, R. G., & Menzies, R. E. (2014). Death
anxiety and its role in psychopathology: Reviewing the
status of a transdiagnostic construct. Clinical Psychology
Review, 34, 580–593. doi:10.1016/j.cpr.2014.09.002

John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The
Big Five Inventory–Versions 4a and 54. Berkeley, CA:
University of California, Berkeley, Institute of Personality
and Social Research.

John, O. P., Naumann, L. P., & Soto, C. J. (2008). Paradigm
shift to the integrative Big Five trait taxonomy: History,
measurement, and conceptual issues. In O. P. John, R.
W. Robins, & L. A. Pervin (Eds.), Handbook of personal-
ity: Theory and research (pp. 114–158). New York, NY:
Guilford Press.

Kastenbaum, R., & Heflick, N. A. (2011). Sad to say: Is it
time for sorrow management theory?. Omega – Journal of
Death and Dying, 62, 305–327. doi:10.2190/OM.62.4.a

Kim, S., Thibodeau, R., & Jorgensen, R. S. (2011). Shame,
guilt, and depressive symptoms: A meta-analytic review.
Psychological Bulletin, 137, 68–96. doi:10.1037/a0021466

Kline, R. (2015). The mediation myth. Basic and Applied
Social Psychology, 37, 202–213. doi:10.1080/
01973533.2015.1049349

Kuhn, T. S. (1962). The structure of scientific revolutions.
Chicago: University of Chicago Press.

Lavelock, C. R., Griffin, B. J., & Worthington, E. L. Jr.
(2013). Forgiveness, religiousness, spirituality, and health
in people with physical challenges: A review and a model.
In R. L. Piedmont & A. Village (Eds.), Research in the
social scientific study of religion (Vol. 24, pp. 53–92).
Boston: Brill.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and
coping. New York: Springer.

Marschall, D., Sanfter, J., & Tangney, J. P. (1994). The state
shame and guilt scale. Fairfax, VA: George Mason
University.

Martens, A., Burke, B. L., Schimel, J., & Faucher, E. H.
(2011). Same but different: Meta-analytically examining
the uniqueness of mortality salience effects. European
Journal of Social Psychology, 41, 6–10. doi:10.1002/
ejsp.767

Martin, L. L., & Erber, R. (2006). Introduction [Special
issue]. Psychological Inquiry, 17, 277. doi:10.1080/
10478400701366928

Martin, L. L., & van den Bos, K. (2014). Beyond terror:
Towards a paradigm shift in the study of threat and cul-
ture. European Review of Social Psychology, 25, 32–70.
doi:10.1080/10463283.2014.923144

McAuley, E., Duncan, T. E., & Russell, D. W. (1992).
Measuring causal attributions: The revised causal dimen-
sion scale (CDS-II). Personality and Social Psychology
Bulletin, 18, 566–573. doi:10.1177/0146167292185006

McConnell, J. M. (2015). A conceptual-theoretical-empirical
framework for self-forgiveness: Implications for research
and practice. Basic and Applied Social Psychology, 37,
143–164. doi:/10.1080/01973533.2015.1016160

McConnell, J. M., Dixon, D. N., & Finch, W. H. (2012). An
alternative model of self- forgiveness. The New School
Psychology Bulletin, 9, 35–51.

BASIC AND APPLIED SOCIAL PSYCHOLOGY 371

https://doi.org/10.1037/0022-3514.67.4.627

https://doi.org/10.1037/0022-3514.63.6.913

https://doi.org/10.1521/jscp.2005.24.5.621

https://doi.org/10.1521/jscp.2005.24.5.621

https://doi.org/10.1521/jscp.2008.27.2.174

https://doi.org/10.1089/eco.2012.0070

https://doi.org/10.1177/1745691613506018

https://doi.org/10.1016/j.jesp.2017.10.005

https://doi.org/10.1016/j.jesp.2017.10.005

https://doi.org/10.1037/a0020524

https://doi.org/10.1037/a0020524

https://doi.org/10.1080/15298868.2015.1036919

https://doi.org/10.2190/96X8-FLUT-TCLH-EL71

https://doi.org/10.1207/s15327957pspr1002_1

https://doi.org/10.1002/ejsp.110

https://doi.org/10.1016/j.cpr.2014.09.002

https://doi.org/10.2190/OM.62.4.a

https://doi.org/10.1037/a0021466

https://doi.org/10.1080/01973533.2015.1049349

https://doi.org/10.1080/01973533.2015.1049349

https://doi.org/10.1002/ejsp.767

https://doi.org/10.1002/ejsp.767

https://doi.org/10.1080/10478400701366928

https://doi.org/10.1080/10478400701366928

https://doi.org/10.1080/10463283.2014.923144

https://doi.org/10.1177/0146167292185006

https://doi.org//10.1080/01973533.2015.1016160

McCullough, M. E., Rachal, K. C., Sandage, S. J.,
Worthington, E. L., Jr., Brown, S. W., & Hight, T. L.
(1998). Interpersonal forgiving in close relationships: II.
Theoretical elaboration and measurement. Journal of
Personality and Social Psychology, 75, 1586–1603. doi:
10.1037/0022-3514.75.6.1586

McCullough, M. E., Worthington, E. L., Jr., & Rachal, K. C.
(1997). Interpersonal forgiving in close relationships.
Journal of Personality and Social Psychology, 73, 321–336.
doi:10.1037/0022-3514.73.2.321

McGregor, H., Lieberman, J. D., Solomon, S., Greenberg, J.,
Arndt, J., Simon, L., & Pyszczynski, T. (1998). Terror
management and aggression: Evidence that mortality sali-
ence motivates aggression against worldview threatening
others. Journal of Personality and Social Psychology, 74,
590–605. doi:10.1037/0022-3514.74.3.590

Michelson, A. A., & Morley, E. W. (1887). On the relative
motion of the earth and the luminiferous ether. American
Journal of Science, s3–34, 333–345. doi:10.2475/ajs.s3-
34.203.333

Mikulincer, M., & Florian, V. (2002). The effect of mortality
salience on self-serving attributions—Evidence for the
function of self-esteem as a terror management mechan-
ism. Basic and Applied Social Psychology, 24, 261–271.
doi:10.1207/S15324834BASP2404_2

Motyl, M., Hart, J., Cooper, D. P., Heflick, N., Goldenberg,
J., & Pyszczynski, T. (2013). Creatureliness priming
reduces aggression and support for war. British Journal of
Social Psychology, 52, 648–666. doi:10.1111/j.2044-
8309.2012.02115.x

Mullet, E., Neto, F., & Rivi�ere, S. (2005). Personality and
its effects on revenge, resentment, forgiveness, and
self-forgiveness. In E. L. Worthington Jr. (Ed.), Handbook
of forgiveness (pp. 159–181). New York: Routledge.

Olshtain, E. (1989). Apologies across languages. In S. Blum-
Kulka, J. House, & G. Kasper (Eds.), Cross-cultural prag-
matics: Requests and apologies (pp. 157). Norwood, NJ:
Ablex.

Pope, A. (1796). A select collection of poems. New London:
Springer. (Original work published 1717)

Popper, K. (1963). Conjectures and refutations: The growth
of scientific knowledge. New York, NY: Routledge &
Kegan Paul.

Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007).
Addressing moderated mediation hypotheses: Theory,
methods, and prescriptions. Multivariate Behavioral
Research, 42, 185–227. doi:10.1080/00273170701341316.

Proulx, T. (2012). Threat-compensation in social psych-
ology: Is there a core motivation? [Special issue]. Social
Cognition, 30, 643–651. doi:10.1521/soco.2012.30.6.643

Proulx, T., & Heine, S. J. (2008). The case of the transmog-
rifying experimenter: Affirmation of a moral schema fol-
lowing implicit change detection. Psychological Science,
19, 1294–1300. doi:10.1111/j.1467-9280.2008.02238.x

Proulx, T., Inzlicht, M., & Harmon-Jones, E. (2012).
Understanding all inconsistency compensation as a pallia-
tive response to violated expectations. Trends in Cognitive
Sciences, 16, 285–291. doi:10.1016/j.tics.2012.04.002

Pyszczynski, T., Greenberg, J., & Solomon, S. (1997). Why
do we need what we need? A terror management per-
spective on the roots of human social motivation [Special

issue]. Psychological Inquiry, 8, 1–20. doi:10.1207/
s15327965pli0801_1

Pyszczynski, T., Greenberg, J., & Solomon, S. (1999). A
dual-process model of defense against conscious and
unconscious death-related thoughts: An extension of ter-
ror management theory. Psychological Review, 106,
835–845. doi:10.1037/0033-295X.106.4.835

Pyszczynski, T., Greenberg, J., Solomon, S., Arndt, J., &
Schimel, J. (2004). Why do people need self-esteem? A
theoretical and empirical review. Psychological Bulletin,
130, 435–468. doi:10.1037/0033-2909.130.3.435

Pyszczynski, T., Solomon, S., & Greenberg, J. (2015). Thirty
years of terror management theory: From genesis to reve-
lation. Advances in Experimental Social Psychology, 52,
1–70. doi:10.1016/bs.aesp.2015.03.001

Rammstedt, B., & John, O. P. (2007). Measuring personality
in one minute or less: A 10-item short version of the Big
Five Inventory in English and German. Journal of
Research in Personality, 41, 203–212. doi:10.1016/
j.jrp.2006.02.001

Raskin, R., & Terry, H. (1988). Principal-components ana-
lysis of the Narcissistic Personality Inventory and further
evidence of its construct validity. Journal of Personality
and Social Psychology, 54, 890–902. doi:10.1037/0022-
3514.54.5.890

Robins, R. W., Hendin, H. M., & Trzesniewski, K. H.
(2001). Measuring global self-esteem: Construct valid-
ation of a single-item measure and the Rosenberg self-
esteem scale. Personality and Social Psychology Bulletin,
27, 151–161. doi:10.1177/0146167201272002

Rosenberg, M. (1965). Society and the adolescent self-image.
Princeton: Princeton University Press.

Rosenblatt, A., Greenberg, J., Solomon, S., Pyszczynski, T.,
& Lyon, D. (1989). Evidence for terror management the-
ory: I. The effects of mortality salience on reactions to
those who violate or uphold cultural values. Journal of
Personality and Social Psychology, 57, 681–690.

Rubin, D. B. (1987). Multiple imputation for nonresponse in
surveys. New York: John Wiley & Sons.

Schimel, J., Landau, M., & Hayes, J. (2008). Self-esteem: A
human solution to the problem of death. Social and
Personality Psychology Compass, 2, 1218–1234. doi:
10.1111/j.1751-9004.2008.00104.x

Schimel, J., Wohl, M. J. A., & Williams, T. (2006). Terror
management and trait empathy: Evidence that mortality
salience promotes reactions of forgiveness among people
with high (vs. low) trait empathy. Motivation and
Emotion, 30, 214–227. doi:10.1007/s11031-006-9040-y

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002).
Experimental and quasi-experimental designs for general-
ized causal inference. Boston: Houghton Mifflin.

Shan, J., & Xu, G. (2008). Self-forgiveness–A new way of
mental counseling. Chinese Journal of Clinical Psychology,
16, 92–94.

Sobel, M. E. (1982). Asymptotic confidence intervals for
indirect effects in structural equation models. In S.
Leinhardt (Ed.), Sociological Methodology (pp. 290–312).
Washington, DC: American Sociological Association.

Solomon, S., Greenberg, J., & Pyszczynski, T. (2004). The
cultural animal: Twenty years of terror management the-
ory and research. In J. Greenberg, S. L. Koole, & T.

372 J. M. MCCONNELL

https://doi.org/10.1037/0022-3514.75.6.1586

https://doi.org/10.1037/0022-3514.73.2.321

https://doi.org/10.1037/0022-3514.74.3.590

https://doi.org/10.2475/ajs.s3-34.203.333

https://doi.org/10.2475/ajs.s3-34.203.333

https://doi.org/10.1207/S15324834BASP2404_2

https://doi.org/10.1111/j.2044-8309.2012.02115.x

https://doi.org/10.1111/j.2044-8309.2012.02115.x

https://doi.org/10.1080/00273170701341316.

https://doi.org/10.1521/soco.2012.30.6.643

https://doi.org/10.1111/j.1467-9280.2008.02238.x

https://doi.org/10.1016/j.tics.2012.04.002

https://doi.org/10.1207/s15327965pli0801_1

https://doi.org/10.1207/s15327965pli0801_1

https://doi.org/10.1037/0033-295X.106.4.835

https://doi.org/10.1037/0033-2909.130.3.435

https://doi.org/10.1016/bs.aesp.2015.03.001

https://doi.org/10.1016/j.jrp.2006.02.001

https://doi.org/10.1016/j.jrp.2006.02.001

https://doi.org/10.1037/0022-3514.54.5.890

https://doi.org/10.1037/0022-3514.54.5.890

https://doi.org/10.1177/0146167201272002

https://doi.org/10.1111/j.1751-9004.2008.00104.x

https://doi.org/10.1007/s11031-006-9040-y

Pyszczynski (Eds.), Handbook of Experimental Existential
Psychology (pp. 13–34). New York, NY: Guilford Press.

Solomon, S., Greenberg, J., & Pyszczynski, T. (1997). Return
of the living dead. Psychological Inquiry, 8, 59–71. doi:
10.1207/s15327965pli0801_13

Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005).
Establishing a causal chain: Why experiments are often
more effective than meditational analyses in examining
psychological processes. Journal of Personality and Social
Psychology, 89, 845–851. doi:10.1037/0022-3514.89.6.845

Steinman, C. T., & Updegraff, J. A. (2015). Delay and
death-thought accessibility: A meta-analysis. Personality
and Social Psychology Bulletin, 41, 1682–1696. doi:
10.1177/0146167215607843

Tangney, J. P., Boone, A. L., & Dearing, R. L. (2005).
Forgiving the self: Conceptual issues and empirical find-
ings. In E. L. Worthington, Jr. (Ed.), Handbook of forgive-
ness (pp.143–158). New York, NY: Routledge.

Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt.
New York, NY: The Guilford Press.

Tate, C. U. (2015). On the overuse and misuse of mediation
analysis: It may be a matter of timing. Basic and Applied
Social Psychology, 37, 235–246. doi:10.1080/
01973533.2015.1062380

Thompson, L. Y., Snyder, C. R., Hoffman, L., Michael, S.
T., Rasmussen, H. N., Billings, L. S., … Roberts, D. E.
(2005). Dispositional forgiveness of self, others, and situa-
tions. Journal of Personality, 73, 313–360. doi:10.1111/
j.1467-6494.2005.00311.x

Tracey, T. J. G., & Glidden-Tracey, C. E. (1999). Integration
of theory, research design, measurement, and analysis:
Toward a reasoned argument. The Counseling
Psychologist, 27, 299–324. doi:10.1177/0011000099273002

Trafimow, D. (2012). The role of auxiliary assumption for
the validity of manipulations and measures. Theory &
Psychology, 22, 486–498. doi:10.1177/0959354311429996

Trafimow, D. (2012). The role of mechanisms, integration,
and unification in science and psychology. Theory &
Psychology, 22, 697–703. doi:10.1177/0959354311433929

Trafimow, D. (2014). Editorial. Basic and Applied Social
Psychology, 36, 1–2. doi:10.1080/01973533.2014.865505

Trafimow, D. (2015). Introduction to the special issue on
mediation analyses: What if planetary scientists used
mediation analysis to infer causation? [Special issue].
Basic and Applied Social Psychology, 37, 197–201. doi:
10.1080/01973533.2015.1064290

Trafimow, D., & Hughes, J. S. (2012). Testing the death
thought suppression and rebound hypothesis: Death
thought accessibility following mortality salience
decreases during a delay. Social Psychological and
Personality Science, 3, 622–629. doi:10.1177/
1948550611432938

Vail, K. I., Arndt, J., Motyl, M., & Pyszczynski, T. (2012).
The aftermath of destruction: Images of destroyed build-
ings increase support for war, dogmatism, and death
thought accessibility. Journal of Experimental Social
Psychology, 48, 1069–1081. doi:10.1016/j.jesp.2012.05.004

Vail, K. E., III, Juhl, J., Arndt, J., Vess, M., Routledge, C., &
Rutjens, B. T. (2012). When death is good for life:
Considering the positive trajectories of terror

management. Personality and Social Psychology Review,
16, 303–329. doi:10.1177/1088868312440046

vanDellen, M. R., Campbell, W. K., Hoyle, R. H., &
Bradfield, E. K. (2011). Compensating, resisting, and
breaking: A meta-analytic examination of reactions to
self-esteem threat. Personality and Social Psychology
Review, 15, 51–74. doi:10.1177/1088868310372950

Van Tongeren, D. R., Green, J. D., Davis, D. E.,
Worthington, E. L., & Reid, C. A. (2013). Till death do
us part: Terror management and forgiveness in close rela-
tionships. Personal Relationships, 20, 755–768. doi:
10.1111/pere.12013

Watson, D., Clark, L. E., & Tellegen, A. (1988).
Development and validation of brief measures of positive
and negative affect: The PANAS scales. Journal of
Personality and Social Psychology, 54, 1063–1070. doi:
10.1037/0022-3514.54.6.1063

Webb, J. R., Hirsch, J. K., & Toussaint, L. (2011).
Forgiveness and alcohol problems: A review of the litera-
ture and a call for intervention-based research.
Alcoholism Treatment Quarterly, 29, 245–273. doi:
10.1080/07347324.2011.585922

Webber, D., Zhang, R., Schimel, J., & Blatter, J. (2016).
Finding death in meaninglessness: Evidence that death-
thought accessibility increases in response to meaning
threats. British Journal of Social Psychology, 55, 144–161.
doi:10.1111/bjso.12118

Weiner, B., Graham, S., Peter, O., & Zmuidinas, M. (1991).
Public confession and forgiveness. Journal of Personality,
59, 281–312. doi:10.1111/j.1467-6494.1991.tb00777.x

Wilson, K. M., & Bernas, R. (2011). A good man is hard to
find: Forgiveness, terror management, and religiosity.
Research in the Social Scientific Study of Religion, 22,
126–140. doi:10.1163/ej.9789004207271.i-360.18

Wohl, M. J. A., DeShea, L., & Wahkinney, R. L. (2008).
Looking within: Measuring state self-forgiveness and its rela-
tionship to psychological wellbeing. Canadian Journal of
Behavioural Science, 40, 1–10. doi:10.1037/0008-400x.40.1.1.1

Wohl, M. J. A., & McLaughlin, K. J. (2014). Self-forgiveness:
The good, the bad, and the ugly. Social and Personality
Compass, 8, 422–425. doi:10.1111/spc3.12119

Woodyatt, L., Worthington, E. L. Jr., Wenzel, M., & Griffin,
B. J. (Eds.). (2017). Handbook of the psychology of self-for-
giveness. Cham, Switzerland: Springer International
Publishing AG.

Worthington, E. L., Jr., Witvliet, C. V. O., Pietrini, P., &
Miller, A. J. (2007). Forgiveness, health, and well-being:
Review of evidence for emotional versus decisional for-
giveness, dispositional forgiveness, and reduced unfor-
giveness. Journal of Behavioral Medicine, 30, 291–302.
doi:10.1007/s10865-007-9105-8

Yen, C. L., & Cheng, C. P. (2010). Terror management
among Taiwanese: Worldview defence or resigning to
fate? Asian Journal of Social Psychology, 13, 185–194. doi:
10.1111/j.1467- 839X.2010.01328.x

Yen, C. L., & Cheng, C. P. (2013). Researcher effects on
mortality salience research: A meta- analytic moderator
analysis. Death Studies, 37, 636–652. doi:10.1080/
07481187.2012.682290

BASIC AND APPLIED SOCIAL PSYCHOLOGY 373

https://doi.org/10.1207/s15327965pli0801_13

https://doi.org/10.1037/0022-3514.89.6.845

https://doi.org/10.1177/0146167215607843

https://doi.org/10.1080/01973533.2015.1062380

https://doi.org/10.1080/01973533.2015.1062380

https://doi.org/10.1111/j.1467-6494.2005.00311.x

https://doi.org/10.1111/j.1467-6494.2005.00311.x

https://doi.org/10.1177/0011000099273002

https://doi.org/10.1177/0959354311429996

https://doi.org/10.1177/0959354311433929

https://doi.org/10.1080/01973533.2014.865505

https://doi.org/10.1080/01973533.2015.1064290

https://doi.org/10.1177/1948550611432938

https://doi.org/10.1177/1948550611432938

https://doi.org/10.1016/j.jesp.2012.05.004

https://doi.org/10.1177/1088868312440046

https://doi.org/10.1177/1088868310372950

https://doi.org/10.1111/pere.12013

https://doi.org/10.1037/0022-3514.54.6.1063

https://doi.org/10.1080/07347324.2011.585922

https://doi.org/10.1111/bjso.12118

https://doi.org/10.1111/j.1467-6494.1991.tb00777.x

https://doi.org/10.1163/ej.9789004207271.i-360.18

https://doi.org/10.1037/0008-400x.40.1.1.1

https://doi.org/10.1111/spc3.12119

https://doi.org/10.1007/s10865-007-9105-8

https://doi.org/10.1111/j.1467- 839X.2010.01328.x

https://doi.org/10.1080/07481187.2012.682290

https://doi.org/10.1080/07481187.2012.682290

Copyright of Basic & Applied Social Psychology is the property of Taylor & Francis Ltd and
its content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder’s express written permission. However, users may print, download, or email
articles for individual use.

• Abstract

Psychological origins of self-forgiveness
Terror management theory
Mortality salience hypothesis
Death-thought accessibility hypothesis
Anxiety-buffer hypothesis
Hypotheses summary
Self-forgiveness: A self-esteem distal defense process?
General method
Experiment 1
Method
Participants
Procedure
Materials
Offense packet
Death awareness packet
Distractor-delay measures
Mediating measure: DTA
Deception
Criterion measures
Additional moderating measure: Effort
Demographic questionnaire
Results and discussion
Missing data
Responsibility
Intercorrelation matrices
Main analyses
Mediation hypothesis
Moderated-mediation hypothesis
Interactional-moderated-mediation hypothesis
Summary

Experiment 2
Method
Participants
Procedure
Materials
Offense packet
Death awareness packet
Results and discussion
Manipulation check
Main analyses
Exploratory analyses
Summary

Experiment 3
Method
Participants
Procedure
Materials
Neutral word-fragment completion task
Trait Self-Forgiveness
Results and discussion
Main analyses
Exploratory analysis
Summary

Experiment 4
Method
Participants
Procedure and materials
Results and discussion
Manipulation check
Main analyses
Exploratory analyses
Summary

General discussion
Review of findings
Eternal sunshine of the spotless mind or spotty empirical findings?
Strengths
Limitations
Conclusion
Acknowledgments
References